Spaces:
Sleeping
Sleeping
File size: 136,155 Bytes
ae39be6 53695a0 69f77f3 53695a0 69f77f3 53695a0 ae39be6 9f24dd6 ae39be6 69f77f3 0697595 9f24dd6 ae39be6 0697595 69f77f3 ae39be6 53695a0 9f24dd6 53695a0 c8e2b80 ae39be6 53695a0 0697595 53695a0 0697595 c8e2b80 53695a0 0697595 53695a0 c8e2b80 53695a0 0697595 53695a0 0697595 69f77f3 53695a0 0697595 53695a0 a517791 69f77f3 53695a0 0697595 99e7c06 0697595 53695a0 0697595 53695a0 aa4ee26 0697595 53695a0 69f77f3 53695a0 a517791 53695a0 ae39be6 53695a0 ae39be6 53695a0 a517791 6f88363 a517791 0697595 aa4ee26 0697595 69f77f3 53695a0 69f77f3 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 69f77f3 53695a0 0697595 53695a0 0697595 a517791 0697595 53695a0 69f77f3 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 69f77f3 53695a0 0697595 53695a0 0697595 a517791 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 9f24dd6 0697595 69f77f3 0697595 69f77f3 0697595 a517791 0697595 69f77f3 9f24dd6 69f77f3 0697595 69f77f3 0697595 a517791 0697595 9f24dd6 69f77f3 53695a0 9f24dd6 53695a0 ae39be6 53695a0 ae39be6 263467b ae39be6 53695a0 a517791 53695a0 a517791 53695a0 0697595 53695a0 5d8249f 69f77f3 5d8249f 69f77f3 53695a0 69f77f3 53695a0 a517791 53695a0 69f77f3 53695a0 69f77f3 53695a0 0697595 53695a0 69f77f3 ae39be6 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 53695a0 0697595 69f77f3 53695a0 0697595 53695a0 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 53695a0 69f77f3 ae39be6 5d8249f 69f77f3 53695a0 0697595 a517791 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 0697595 ae39be6 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 0697595 ae39be6 0697595 c8e2b80 0697595 a517791 0697595 a517791 0697595 a517791 0697595 a517791 0697595 c8e2b80 ae39be6 c8e2b80 ae39be6 0697595 53695a0 0697595 aa4ee26 0697595 3fcf7c2 ae39be6 53695a0 0697595 53695a0 ae39be6 69f77f3 53695a0 69f77f3 53695a0 ae39be6 69f77f3 53695a0 69f77f3 0697595 aa4ee26 53695a0 c8e2b80 53695a0 c8e2b80 aa4ee26 c8e2b80 0697595 53695a0 c8e2b80 53695a0 c8e2b80 53695a0 c8e2b80 53695a0 0697595 c8e2b80 0697595 53695a0 0697595 53695a0 ae39be6 53695a0 aa4ee26 6f88363 aa4ee26 c8e2b80 53695a0 0697595 ae39be6 69f77f3 ae39be6 53695a0 0697595 6f88363 ae39be6 53695a0 99e7c06 0697595 aa4ee26 0697595 53695a0 99e7c06 0697595 53695a0 69f77f3 53695a0 69f77f3 53695a0 0697595 53695a0 99e7c06 53695a0 aa4ee26 99e7c06 aa4ee26 0697595 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 0697595 53695a0 ae39be6 53695a0 ae39be6 53695a0 0697595 ae39be6 53695a0 c8e2b80 53695a0 69f77f3 c8e2b80 0697595 c8e2b80 0697595 53695a0 99e7c06 aa4ee26 53695a0 a517791 53695a0 0697595 53695a0 0697595 53695a0 a517791 53695a0 ae39be6 a517791 ae39be6 a517791 53695a0 ae39be6 53695a0 ae39be6 53695a0 0697595 53695a0 a517791 53695a0 0697595 53695a0 0697595 53695a0 69f77f3 c8e2b80 69f77f3 c8e2b80 69f77f3 c8e2b80 69f77f3 0697595 ae39be6 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 9f24dd6 69f77f3 aa4ee26 69f77f3 c8e2b80 69f77f3 c8e2b80 69f77f3 0697595 69f77f3 ae39be6 69f77f3 c8e2b80 69f77f3 ae39be6 69f77f3 0697595 ae39be6 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 0697595 69f77f3 ae39be6 69f77f3 ae39be6 69f77f3 ae39be6 69f77f3 ae39be6 69f77f3 0697595 69f77f3 ae39be6 69f77f3 0697595 69f77f3 0697595 53695a0 a517791 53695a0 69f77f3 53695a0 5d8249f 69f77f3 53695a0 69f77f3 53695a0 a517791 53695a0 69f77f3 53695a0 69f77f3 53695a0 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 0697595 53695a0 c8e2b80 53695a0 0697595 53695a0 0697595 53695a0 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 0697595 c8e2b80 69f77f3 c8e2b80 ae39be6 c8e2b80 ae39be6 53695a0 c8e2b80 53695a0 ae39be6 53695a0 ae39be6 53695a0 ae39be6 53695a0 ae39be6 53695a0 ae39be6 53695a0 aa4ee26 53695a0 ae39be6 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 |
import os, io, re, uuid, json, time, torch, base64
import shutil, zipfile, requests, tempfile, subprocess, threading, contextlib
import numpy as np
import gradio as gr
import pandas as pd
import plotly.express as px
import plotly.graph_objects as go
from yaml import safe_dump, safe_load
from rdkit import Chem, RDLogger
from rdkit.Chem import AllChem, Descriptors
from rdkit.Geometry import Point3D
from rdkit.Chem.rdDetermineBonds import DetermineConnectivity
from rdkit.Contrib.SA_Score import sascorer # type: ignore
from rdkit.Contrib.NP_Score import npscorer # type: ignore
from pathlib import Path
from boltz.main import download_boltz2
from boltz.data import const
from concurrent.futures import ThreadPoolExecutor, as_completed
from functools import partial
from gemmi import cif # type: ignore
# TODO: Convert AF3/Chai-1/Protenix JSON to Boltz YAML
RDLogger.DisableLog('rdApp.*')
with contextlib.redirect_stdout(io.StringIO()), contextlib.redirect_stderr(io.StringIO()):
fscore = npscorer.readNPModel()
periodic_table = Chem.GetPeriodicTable()
entity_types = ['Protein', 'DNA', 'RNA', 'Ligand', 'CCD']
entity_label_map = {'Protein': 'Sequence', 'DNA': 'Sequence', 'RNA': 'Sequence',
'Ligand': 'SMILES', 'CCD': 'CCD Code'}
allow_char_dict = {'Protein': "ACDEFGHIKLMNPQRSTVWY",
'DNA' : "ACGT",
'RNA' : "ACGU"}
rev_comp_map = {'DNA': {'A': 'T', 'T': 'A', 'C': 'G', 'G': 'C', 'U': 'A'},
'RNA': {'A': 'U', 'U': 'A', 'C': 'G', 'G': 'C', 'T': 'A'}}
property_functions = {'Molecular Weight' : Descriptors.MolWt,
'Num. of Hydrogen Bond Donors' : Descriptors.NumHDonors,
'Num. of Hydrogen Bond Acceptors' : Descriptors.NumHAcceptors,
'LogP': Descriptors.MolLogP,
'Topological Polar Surface Area (TPSA)': Descriptors.TPSA,
'Rotatable Bonds' : Descriptors.NumRotatableBonds,
'Num. of Rings' : Descriptors.RingCount,
'Formal Charge' : lambda mol: sum([atom.GetFormalCharge() for atom in mol.GetAtoms()]),
'Num. of Heavy Atoms' : Descriptors.HeavyAtomCount,
'Num. of Atoms' : lambda mol: mol.GetNumAtoms(),
'Molar Refractivity' : Descriptors.MolMR,
'Quantitative Estimate of Drug-Likeness (QED)' : Descriptors.qed,
'Natural Product-likeness Score (NP)': partial(npscorer.scoreMol, fscore=fscore),
'Synthetic Accessibility Score (SA)': sascorer.calculateScore}
css = """
footer { display: none !important; }
.sequence textarea {font-family: Courier New, Courier, monospace; !important}
.validation {font-size: 12px; font-family: Courier New, Courier, monospace; !important}
.log textarea {font-size: 12px; font-family: Courier New, Courier, monospace; !important}
.small-upload-style .wrap {font-size: 10px; !important}
.small-upload-style .icon-wrap svg {display: none; !important}
"""
device_num = 1
cuda_available = torch.cuda.is_available()
if cuda_available:
device_num = torch.cuda.device_count()
curr_dir = os.path.dirname(__file__)
output_dir = os.path.join(curr_dir, 'boltz_output')
# output_dir = os.path.join(curr_dir, 'boltz_vhts')
os.makedirs(output_dir, exist_ok=True)
template_dir = os.path.join(curr_dir, 'templates')
os.makedirs(template_dir, exist_ok=True)
msa_dir = os.path.join(curr_dir, 'usr_msa')
os.makedirs(msa_dir, exist_ok=True)
input_dir = os.path.join(curr_dir, 'boltz_input')
os.makedirs(input_dir, exist_ok=True)
shutil.rmtree(input_dir)
os.makedirs(input_dir)
### Defining Boltz parameters ###
device_number = gr.Number(1, label='devices',
info='The number of devices to use for prediction.',
minimum=1, maximum=device_num, interactive=True)
acc_choices = ['gpu', 'cpu'] if cuda_available else ['cpu']
accelerator_type = gr.Dropdown(acc_choices, value=acc_choices[0],
info='The accelerator to use for prediction.',
label='accelerator', interactive=True)
recycling_steps = gr.Number(3, label='recycling_steps',
info='The number of recycling steps to use for prediction.',
minimum=1, interactive=True)
sampling_steps = gr.Number(200, label='sampling_steps',
info='The number of sampling steps to use for prediction.',
minimum=1, interactive=True)
diffusion_samples = gr.Number(3, label='diffusion_samples',
info='The number of diffusion samples to use for prediction.',
minimum=1, interactive=True)
step_scale = gr.Number(1.638, label='step_scale',
info='The lower the higher the diversity among samples (recommended between 1 and 2).',
minimum=0., interactive=True)
num_workers = gr.Number(2, label='num_workers', minimum=0,
info='The number of dataloader workers to use for prediction.',
maximum=os.cpu_count(), interactive=True)
preprocessing_threads = gr.Number(os.cpu_count(), label='preprocessing-threads',
info='The number of threads to use for preprocessing.',
interactive=True, minimum=1, maximum=os.cpu_count())
affinity_mw_correction = gr.Checkbox(False, label='affinity_mw_correction',
info='Whether to add the Molecular Weight correction to the affinity value head.',
interactive=True,)
sampling_steps_affinity = gr.Number(200, label='sampling_steps_affinity',
info='The number of sampling steps to use for affinity prediction.',
interactive=True, minimum=1)
diffusion_samples_affinity = gr.Number(5, label='diffusion_samples_affinity',
info='The number of diffusion samples to use for affinity prediction.',
interactive=True, minimum=1)
no_trifast = gr.Checkbox(False if cuda_available else True, label='no_trifast',
info='Whether to NOT use trifast kernels for triangular updates.')
override = gr.Checkbox(False, label='override', info='Whether to override existing predictions if found.')
use_potentials = gr.Checkbox(False, label='use_potentials',
info='Whether to run the original Boltz-2 model using inference time potentials.')
boltz_method = gr.Dropdown(list(const.method_types_ids.keys()), label='method',
value='x-ray diffraction',
info='The method to use for prediction.')
all_boltz_parameters = [device_number, accelerator_type, recycling_steps, sampling_steps,
diffusion_samples, step_scale, num_workers, preprocessing_threads,
affinity_mw_correction, sampling_steps_affinity, diffusion_samples_affinity,
use_potentials, boltz_method, no_trifast, override]
def concurrent_download_model_weight():
cache_pth = Path('~/.boltz').expanduser()
cache_pth.mkdir(exist_ok=True)
all_files = os.listdir(cache_pth)
if ('mols' in all_files and 'ccd.pkl' in all_files and
'boltz2_conf.ckpt' in all_files and 'boltz2_aff.ckpt' in all_files):
return
download_boltz2(cache_pth)
return
def manual_download_boltz_weights():
cache_pth = Path('~/.boltz').expanduser()
cache_pth.mkdir(exist_ok=True)
all_files = os.listdir(cache_pth)
if ('mols' in all_files and 'ccd.pkl' in all_files and
'boltz2_conf.ckpt' in all_files and 'boltz2_aff.ckpt' in all_files):
yield gr.update(interactive=True, value='Weight downloaded!')
yield gr.update(interactive=False, value='Downloading...')
download_boltz2(cache_pth)
yield gr.update(interactive=True, value='Weight downloaded!')
### Boltz parameters end ###
def check_dir_exist_and_rename(dir_pth: str):
basename = os.path.basename(dir_pth).rsplit('_', 1)[0]
dirname = os.path.dirname(dir_pth)
while os.path.isdir(dir_pth):
dir_pth = os.path.join(dirname, f'{basename}_{uuid.uuid4().hex[:8]}')
os.makedirs(dir_pth)
def _check_yaml_strings(yaml_str: str):
if not yaml_str:
return False
yaml_dict = safe_load(yaml_str)
if 'sequences' not in yaml_dict or len(yaml_dict['sequences']) < 1:
return False
for seq_dict in yaml_dict['sequences']:
k = list(seq_dict.keys())[0]
if k not in ['protein', 'ligand', 'rna', 'dna'] or len(seq_dict) > 1:
return False
seq_info_dict = seq_dict[k]
if 'id' not in seq_info_dict or ('sequence' not in seq_info_dict and
'smiles' not in seq_info_dict and
'ccd' not in seq_info_dict):
return False
return True
def check_yaml_strings(yaml_str: str, *args):
final_bool_args = []
for value in args:
if isinstance(value, pd.DataFrame):
final_bool_args.append(not value.empty)
else:
final_bool_args.append(value)
return gr.update(interactive=_check_yaml_strings(yaml_str) & all(final_bool_args))
def check_batch_yaml_and_name(yaml_str: str, name_str: str):
name_valid = bool(name_str.strip())
yaml_valid = _check_yaml_strings(yaml_str)
validity_text = ''
if not name_valid:
validity_text += 'Missing name. '
if not yaml_valid:
validity_text += 'Invalid yaml file.'
return gr.update(info=validity_text)
def clear_curr_batch_dict():
return {}, 0
def upload_multi_files(files: list[str], curr_cnt: int):
final_yaml_dict = {}
for file in files:
base_name = os.path.basename(file).rsplit('.', 1)[0]
with open(file) as f:
yaml_str = f.read()
if _check_yaml_strings(yaml_str):
final_yaml_dict[base_name] = yaml_str
os.remove(file)
curr_cnt += len(final_yaml_dict)
return final_yaml_dict, curr_cnt, None
def add_current_single_to_batch(name: str, yaml_str: str, curr_yaml_dict: dict, curr_cnt: int):
if name in curr_yaml_dict:
i = 2
new_name = f'{name}_{i}'
while new_name in curr_yaml_dict:
i += 1
new_name = f'{name}_{i}'
name = new_name
curr_yaml_dict[name] = yaml_str
yield curr_yaml_dict, curr_cnt + 1, 'Complex added!'
time.sleep(2.)
yield gr.update(), gr.update(), 'Add to Batch'
def read_tempaltes(files: list[str], old_cif_name_chain_dict: dict,
old_cif_name_path_dict: dict, old_usage_dict: dict,
old_template_name_setting_dict: dict):
if not old_cif_name_path_dict:
saved_cif_dir = os.path.join(template_dir, uuid.uuid4().hex[:8])
check_dir_exist_and_rename(saved_cif_dir)
else:
written_file = list(old_cif_name_path_dict.values())[0]
saved_cif_dir = os.path.dirname(written_file)
for cif_file in files:
name = os.path.basename(cif_file).rsplit('.', 1)[0]
new_template_pth = os.path.join(saved_cif_dir, os.path.basename(cif_file))
chain_index = 0
stop_search_chain = False
unique_chains = set()
with open(cif_file) as f:
cif_str = f.read()
with open(new_template_pth, 'w') as f:
f.write(cif_str)
for line in cif_str.splitlines():
if line.startswith('_atom_site.') and not stop_search_chain:
label = line.strip().split('_atom_site.', 1)[-1]
if label == 'label_asym_id':
stop_search_chain = True
else:
chain_index += 1
elif line.startswith(('HETATM', 'ATOM')):
chain = line.split()[chain_index]
unique_chains.update(chain)
elif stop_search_chain and line.strip() == '#':
break
old_cif_name_chain_dict[name] = sorted(list(unique_chains))
old_cif_name_path_dict[name] = new_template_pth
old_template_name_setting_dict[name] = {'chain_id': [], 'template_id': []}
if name not in old_usage_dict:
old_usage_dict[name] = True
return (gr.update(choices=list(old_cif_name_chain_dict), value=list(old_cif_name_chain_dict)[0]),
old_cif_name_chain_dict, old_cif_name_path_dict, old_usage_dict,
gr.update(interactive=bool(old_usage_dict), value=old_usage_dict[list(old_cif_name_chain_dict)[0]]),
old_template_name_setting_dict)
def update_template_chain_ids_and_settings(curr_usage_bool: bool, target_chain_ids: list, template_chain_ids: list,
curr_name: str,
template_name_usage_dict: dict, template_name_setting_dict: dict):
template_name_usage_dict[curr_name] = curr_usage_bool
template_name_setting_dict[curr_name]['chain_id'] = target_chain_ids
template_name_setting_dict[curr_name]['template_id'] = template_chain_ids
return template_name_usage_dict, template_name_setting_dict
def update_template_dropdown(curr_name: str, template_name_chain_dict: dict,
template_name_usage_dict: dict, template_name_setting_dict: dict):
return (template_name_usage_dict[curr_name],
template_name_setting_dict[curr_name]['chain_id'],
gr.update(value=template_name_setting_dict[curr_name]['template_id'],
choices=template_name_chain_dict[curr_name]))
def update_bond_sequence_length_with_chain(sel_chain: str, mapping_dict: dict):
data_dict = mapping_dict.get(sel_chain, None)
if data_dict is None:
return gr.update(choices=None, value=None)
if data_dict['type'] in ['CCD']:
return gr.update(choices=['1'], value='1', interactive=True)
elif data_dict['type'] in ['Protein', 'DNA', 'RNA']:
total_len = len(data_dict['sequence'])
return gr.update(choices=[str(i) for i in range(1, total_len+1)], value='1', interactive=True)
else:
return gr.update(choices=None, value=None)
### Running Boltz ###
def execute_single_boltz(file_name: str, yaml_str: str,
devices: int, accelerator: str,
recycling_steps: int, sampling_steps: int,
diffusion_samples: float, step_scale: int,
num_workers: int, preprocessing_threads: int,
affinity_mw_correction: bool,
sampling_steps_affinity: int, diffusion_samples_affinity: int,
use_potentials: bool, boltz_method: str, no_trifast: bool, override: bool):
random_dir_name = f"{file_name}_{uuid.uuid4().hex[:8]}"
inp_rng_dir = os.path.join(input_dir, random_dir_name)
out_rng_dir = os.path.join(output_dir, random_dir_name)
check_dir_exist_and_rename(inp_rng_dir)
check_dir_exist_and_rename(out_rng_dir)
inp_yaml = os.path.join(inp_rng_dir, file_name+'.yaml')
with open(inp_yaml, 'w') as f:
f.write(yaml_str)
final_strs = ['--use_msa_server', '--write_full_pae', '--write_full_pde']
if use_potentials:
final_strs.append('--use_potentials')
if affinity_mw_correction:
final_strs.append('--affinity_mw_correction')
if no_trifast:
final_strs.append('--no_trifast')
if override:
final_strs.append('--override')
cmd = ['boltz', 'predict', inp_yaml,
'--out_dir', out_rng_dir,
'--devices', str(devices),
'--accelerator', accelerator,
'--recycling_steps', str(recycling_steps),
'--sampling_steps', str(sampling_steps),
'--diffusion_samples', str(diffusion_samples),
'--step_scale', str(step_scale),
'--num_workers', str(num_workers),
'--preprocessing-threads', str(preprocessing_threads),
'--sampling_steps_affinity', str(sampling_steps_affinity),
'--diffusion_samples_affinity', str(diffusion_samples_affinity),
'--method', boltz_method]
cmd += final_strs
yield gr.update(value='Predicting...', interactive=False), ''
full_output = ''
curr_running_process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
text=True, encoding="utf-8")
for line in iter(curr_running_process.stdout.readline, ''):
if 'The loaded checkpoint was produced with' in line or\
'You are using a CUDA device' in line: # Just skip these warnings
continue
if line.startswith('Predicting DataLoader'):
full_output = full_output.rsplit('\n', 2)[0] + '\n' + line
else:
full_output += line
yield gr.update(value='Predicting...', interactive=False), full_output
curr_running_process.stdout.close()
curr_running_process.wait()
full_output += 'Prediction Done!\n'
yield gr.update(value='Run Boltz', interactive=True), full_output
def execute_multi_boltz(all_files: list[str],
devices: int, accelerator: str,
recycling_steps: int, sampling_steps: int,
diffusion_samples: float, step_scale: int,
num_workers: int, preprocessing_threads: int,
affinity_mw_correction: bool,
sampling_steps_affinity: int, diffusion_samples_affinity: int,
use_potentials: bool, boltz_method: str, no_trifast: bool, override: bool):
# even though all the files are passed here, only their directory will be used
# since Boltz inherently allow batch processing
dirname = os.path.dirname(all_files[0])
rng_basename = os.path.basename(dirname)
out_rng_dir = os.path.join(output_dir, rng_basename)
check_dir_exist_and_rename(out_rng_dir)
final_strs = ['--use_msa_server', '--write_full_pae', '--write_full_pde']
if use_potentials:
final_strs.append('--use_potentials')
if affinity_mw_correction:
final_strs.append('--affinity_mw_correction')
if no_trifast:
final_strs.append('--no_trifast')
if override:
final_strs.append('--override')
cmd = ['boltz', 'predict', dirname,
'--out_dir', out_rng_dir,
'--devices', str(devices),
'--accelerator', accelerator,
'--recycling_steps', str(recycling_steps),
'--sampling_steps', str(sampling_steps),
'--diffusion_samples', str(diffusion_samples),
'--step_scale', str(step_scale),
'--num_workers', str(num_workers),
'--preprocessing-threads', str(preprocessing_threads),
'--sampling_steps_affinity', str(sampling_steps_affinity),
'--diffusion_samples_affinity', str(diffusion_samples_affinity),
'--method', boltz_method]
cmd += final_strs
yield gr.update(value='Predicting...', interactive=False), ''
full_output = ''
curr_running_process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
text=True, encoding="utf-8")
for line in iter(curr_running_process.stdout.readline, ''):
if 'The loaded checkpoint was produced with' in line or\
'You are using a CUDA device' in line: # Just skip these warnings
continue
if line.startswith('Predicting DataLoader'):
full_output = full_output.rsplit('\n', 2)[0] + '\n' + line
else:
full_output += line
yield gr.update(value='Predicting...', interactive=False), full_output
curr_running_process.stdout.close()
curr_running_process.wait()
full_output += 'Prediction Done!\n'
yield gr.update(value='Batch Predict', interactive=True), full_output
def execute_vhts_boltz(file_prefix: str, all_ligands: pd.DataFrame,
ligand_chain: str, yaml_str: str,
devices: int, accelerator: str,
recycling_steps: int, sampling_steps: int,
diffusion_samples: float, step_scale: int,
num_workers: int, preprocessing_threads: int,
affinity_mw_correction: bool,
sampling_steps_affinity: int, diffusion_samples_affinity: int,
use_potentials: bool, boltz_method: str, no_trifast: bool, override: bool):
random_dir_name = f"{file_prefix}_vHTS_{uuid.uuid4().hex[:8]}"
inp_rng_dir = os.path.join(input_dir, random_dir_name)
out_rng_dir = os.path.join(output_dir, random_dir_name)
check_dir_exist_and_rename(inp_rng_dir)
check_dir_exist_and_rename(out_rng_dir)
yaml_template_dict = safe_load(yaml_str)
final_strs = ['--use_msa_server', '--write_full_pae', '--write_full_pde']
if use_potentials:
final_strs.append('--use_potentials')
if affinity_mw_correction:
final_strs.append('--affinity_mw_correction')
if no_trifast:
final_strs.append('--no_trifast')
# Never override for vHTS
# if override:
# final_strs.append('--override')
cmd = ['boltz', 'predict', inp_rng_dir,
'--out_dir', out_rng_dir,
'--devices', str(devices),
'--accelerator', accelerator,
'--recycling_steps', str(recycling_steps),
'--sampling_steps', str(sampling_steps),
'--diffusion_samples', str(diffusion_samples),
'--step_scale', str(step_scale),
'--num_workers', str(num_workers),
'--preprocessing-threads', str(preprocessing_threads),
'--sampling_steps_affinity', str(sampling_steps_affinity),
'--diffusion_samples_affinity', str(diffusion_samples_affinity),
'--method', boltz_method]
cmd += final_strs
for idx, row in all_ligands.iterrows():
n, s = row['Name'], row['SMILES']
for seq_info in yaml_template_dict['sequences']:
if 'ligand' in seq_info and seq_info['ligand']['id'] == ligand_chain:
seq_info['ligand']['smiles'] = s
break
inp_yaml_pth = os.path.join(inp_rng_dir, f'{n}.yaml')
with open(inp_yaml_pth, 'w') as f:
f.write(safe_dump(yaml_template_dict))
# execute on only a single file to retrieve msa, prevent colabfold server overload
if idx == 0:
yield gr.update(value='Predicting...', interactive=False), ''
full_output = ''
curr_running_process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
text=True, encoding="utf-8")
for line in iter(curr_running_process.stdout.readline, ''):
if 'The loaded checkpoint was produced with' in line or\
'You are using a CUDA device' in line: # Just skip these warnings
continue
if line.startswith('Predicting DataLoader'):
full_output = full_output.rsplit('\n', 2)[0] + '\n' + line
else:
full_output += line
yield gr.update(value='Predicting...', interactive=False), full_output
curr_running_process.stdout.close()
curr_running_process.wait()
num_msa_pth_map = {}
msa_dir = os.path.join(out_rng_dir, f'boltz_results_{random_dir_name}', 'msa')
for msa_f in os.listdir(msa_dir):
if msa_f.endswith('.csv'):
num = msa_f.rsplit('.', 1)[0].rsplit('_', 1)[-1]
num_msa_pth_map[int(num)] = os.path.join(msa_dir, msa_f)
# Just add the csv path containing the MSA to the "msa" key of template.
# Number by the index of list within the "sequences" key!
for seq_num, seq_info in enumerate(yaml_template_dict['sequences']):
if seq_num in num_msa_pth_map:
seq_info['protein']['msa'] = num_msa_pth_map[seq_num]
cmd.remove('--use_msa_server')
curr_running_process = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.STDOUT,
text=True, encoding="utf-8")
for line in iter(curr_running_process.stdout.readline, ''):
if 'The loaded checkpoint was produced with' in line or\
'You are using a CUDA device' in line: # Just skip these warnings
continue
if line.startswith('Predicting DataLoader'):
full_output = full_output.rsplit('\n', 2)[0] + '\n' + line
else:
full_output += line
yield gr.update(value='Predicting...', interactive=False), full_output
curr_running_process.stdout.close()
curr_running_process.wait()
full_output += 'Prediction Done. Processing ligand to SDF format...\n'
out_pred_dir = Path(os.path.join(out_rng_dir, f'boltz_results_{random_dir_name}', 'predictions'))
dir_smiles_dict = {}
for _, row in all_ligands.iterrows():
name, smiles = row['Name'], row['SMILES']
dir_smiles_dict[out_pred_dir / f'{name}'] = smiles
with ThreadPoolExecutor() as executor:
futures = [executor.submit(_recover_dir_molecule, cif_dir, smiles, ligand_chain) for
cif_dir, smiles in dir_smiles_dict.items()]
total = len(futures)
n = 0
errors = ''
progress_text = f'SDF Format Conversion Progress: {n} / {total}'
yield gr.update(value='Predicting...', interactive=False), full_output + progress_text
for f in as_completed(futures):
err = f.result()
if err:
errors += err
n += 1
progress_text = f'SDF Format Conversion Progress: {n} / {total}'
yield gr.update(value='Predicting...', interactive=False), full_output + errors + progress_text
progress_text += '\nvHTS done!'
yield gr.update(value='Run vHTS', interactive=True), full_output + errors + progress_text
### vHTS ###
def update_chem_file_format(chem_type: str):
if chem_type == 'Chemical files':
file_types=['.sdf', '.mol', '.smi', '.zip']
label = 'Upload chemical file(s)'
tabular_visibility = gr.update(visible=False)
elif chem_type == 'Tabular files':
file_types = ['.csv', '.tsv', '.txt']
label = 'Upload tabular file(s)'
tabular_visibility = gr.update(visible=True)
return (gr.update(file_types=file_types, label=label), tabular_visibility,
tabular_visibility, tabular_visibility)
def __check_smi_title_line(smi_file: str):
with open(smi_file) as f:
for r, l in enumerate(f):
possible_smiles = l.split(' ')[0]
if Chem.MolFromSmiles(possible_smiles) is not None:
return r
return 0
def _process_single_chem_file(chem_f: str):
if chem_f.endswith('.sdf'):
mols = Chem.MultithreadedSDMolSupplier(chem_f)
elif chem_f.endswith('.mol'):
mols = [Chem.MolFromMolFile(chem_f)]
elif chem_f.endswith('.smi'):
n = __check_smi_title_line(chem_f)
mols = Chem.MultithreadedSmilesMolSupplier(chem_f, titleLine=n)
names, smiles = [], []
for mol in mols:
if mol is None:
continue
if mol.HasProp('_Name'):
name = mol.GetProp('_Name')
else:
name = os.path.basename(chem_f).rsplit('.', 1)[0]
smi = Chem.MolToSmiles(mol)
names.append(name)
smiles.append(smi)
return names, smiles
def _process_uploaded_chem_file(f: str):
if f.endswith(('.sdf', '.mol', '.smi')):
final_names, final_smiles = _process_single_chem_file(f)
elif f.endswith('.zip'):
with zipfile.ZipFile(f, 'r') as zip_ref:
final_names, final_smiles = [], []
for filename in zip_ref.namelist():
if filename.endswith(('.sdf', '.mol', '.smi', '.zip')):
with zip_ref.open(filename) as file_in_zip:
file_content = file_in_zip.read().decode()
with tempfile.NamedTemporaryFile(suffix='.'+filename.rsplit('.', 1)[-1], delete=False) as temp_file:
temp_file.write(file_content.encode('utf-8'))
temp_file.flush()
temp_file_path = temp_file.name
extracted_n, extracted_s = _process_uploaded_chem_file(temp_file_path)
os.remove(temp_file_path)
final_names.extend(extracted_n)
final_smiles.extend(extracted_s)
return [final_names, final_smiles]
def _process_tabular_files(chem_f: list[str], name_col: str, chem_col: str, delimiter: str):
try:
df = pd.read_csv(chem_f, delimiter=delimiter)
if name_col in df and chem_col in df:
df = df[[name_col, chem_col]].dropna()
else:
return [], []
except:
return [], []
final_names, final_smiles = [], []
for _, row in df.iterrows():
name = row[name_col]
chem_str = row[chem_col]
if chem_str.startswith('InChI='):
mol = Chem.MolFromInchi(chem_str)
else:
mol = Chem.MolFromSmiles(chem_str)
if mol is not None:
smi = Chem.MolToSmiles(mol)
final_names.append(name)
final_smiles.append(smi)
return final_names, final_smiles
def process_uploaded_ligand(chem_files: list[str], name_col: str,
chem_col: str, delimiter: str, prev_df: pd.DataFrame):
total_names, final_smiles = prev_df['Name'].to_list(), prev_df['SMILES'].to_list()
for chem_f in chem_files:
if chem_f.endswith(('.sdf', '.mol', '.smi', '.zip')):
a, b = _process_uploaded_chem_file(chem_f)
elif chem_f.endswith(('.csv', '.tsv', '.txt')):
a, b = _process_tabular_files(chem_f, name_col, chem_col, delimiter)
total_names.extend(a)
final_smiles.extend(b)
final_names = []
for name in total_names:
if name in final_names:
i = 2
new_name = f'{name}_{i}'
while new_name in final_names:
i += 1
new_name = f'{name}_{i}'
name = new_name
final_names.append(name)
return pd.DataFrame({'Name': final_names, 'SMILES': final_smiles})
def __extract_ligand_coord(cif_pth: str, lig_chain: str):
p_map = {'Chain': 0, 'Atom': 0, 'X': 0, 'Y': 0, 'Z': 0}
atom_coord_info = []
n = -1
with open(cif_pth) as f:
for l in f:
if l.startswith('_atom_site.'):
n += 1
if l.startswith('_atom_site.auth_asym_id'):
p_map['Chain'] = n
elif l.startswith('_atom_site.type_symbol'):
p_map['Atom'] = n
elif l.startswith('_atom_site.Cartn_x'):
p_map['X'] = n
elif l.startswith('_atom_site.Cartn_y'):
p_map['Y'] = n
elif l.startswith('_atom_site.Cartn_z'):
p_map['Z'] = n
if l.startswith('HETATM'):
line_splitted = l.split()
if line_splitted[p_map['Chain']] == lig_chain:
a, x, y, z = line_splitted[p_map['Atom']], line_splitted[p_map['X']], \
line_splitted[p_map['Y']], line_splitted[p_map['Z']]
a = Chem.Atom(periodic_table.GetAtomicNumber(a.lower().capitalize()))
atom_coord_info.append((a, Point3D(float(x), float(y), float(z))))
if atom_coord_info and l.startswith('#'):
break
return atom_coord_info
def __reconstruct_mol_from_data(mol_data: list[tuple]):
mol = Chem.EditableMol(Chem.Mol())
conf = Chem.Conformer(len(mol_data))
fc = 0
for i, data in enumerate(mol_data):
atom, coord = data
mol.AddAtom(atom)
conf.SetAtomPosition(i, coord)
fc += atom.GetFormalCharge()
mol = mol.GetMol()
mol.AddConformer(conf)
DetermineConnectivity(mol)
return mol
def _recover_dir_molecule(cif_dir: str, smiles: str, ligand_chain: str):
ref_mol = Chem.MolFromSmiles(smiles)
errors = ''
for f in os.listdir(cif_dir):
if f.endswith('.cif'):
try:
data = __extract_ligand_coord(os.path.join(cif_dir, f), ligand_chain)
coord_mol = __reconstruct_mol_from_data(data)
final_mol = AllChem.AssignBondOrdersFromTemplate(ref_mol, coord_mol)
AllChem.AssignStereochemistryFrom3D(final_mol)
for a in final_mol.GetAtoms():
a.SetNumRadicalElectrons(0)
name = f.rsplit('.', 1)[0]
out_sdf_f = os.path.join(cif_dir, name + '.sdf')
final_mol.SetProp('_Name', name)
final_mol.SetProp('SMILES', Chem.MolToSmiles(final_mol))
with Chem.SDWriter(out_sdf_f) as w:
w.write(final_mol)
except Exception as e:
errors += f'{e}\n'
return errors
### Result visulization ###
def get_molstar_html(mmcif_base64):
return f"""
<iframe
id="molstar_frame"
style="width: 100%; height: 600px; border: none;"
srcdoc='
<!DOCTYPE html>
<html>
<head>
<script src="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.js"></script>
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.css">
</head>
<body>
<div id="Viewer" style="width: 1200px; height: 400px; position: center"></div>
<script>
(async function() {{
const viewer = new rcsbMolstar.Viewer("Viewer",
{{layoutIsExpanded: true,
layoutShowControls: false,
viewportShowExpand: true,
showWelcomeToast: false}});
const mmcifData = "{mmcif_base64}";
const blob = new Blob(
[atob(mmcifData)],
{{ type: "text/plain" }}
);
const url = URL.createObjectURL(blob);
try {{
await viewer.loadStructureFromUrl(url, "mmcif");
}} catch (error) {{
console.error("Error loading structure:", error);
}}
}})();
</script>
</body>
</html>
'>
</iframe>"""
def read_output_files(read_vhts: bool):
name_rank_map = {}
for out_f_or_d in os.listdir(output_dir):
rng_dir = os.path.join(output_dir, out_f_or_d)
is_vhts = out_f_or_d.rsplit('_')[-2] == 'vHTS'
if not read_vhts and is_vhts:
continue
if os.path.isdir(rng_dir):
for target_pth in os.listdir(rng_dir):
if not target_pth.startswith('boltz_results_'):
continue
target_dir = os.path.join(rng_dir, target_pth)
pred_parent_dir = os.path.join(target_dir, 'predictions')
for name in os.listdir(pred_parent_dir):
pred_dir = os.path.join(pred_parent_dir, name)
if not os.path.isdir(pred_dir):
continue
if name in name_rank_map:
i = 2
new_name = f'{name}_{i}'
while new_name in name_rank_map:
i += 1
new_name = f'{name}_{i}'
else:
new_name = name
name_rank_map[new_name] = []
all_files = os.listdir(pred_dir)
total_models = len(all_files) // 5
aff_pth = os.path.join(pred_dir, f'affinity_{name}.json')
if not os.path.exists(aff_pth):
aff_pth = None
for i in range(total_models):
cnf_pth = os.path.join(pred_dir, f'confidence_{name}_model_{i}.json')
mdl_pth = os.path.join(pred_dir, f'{name}_model_{i}.cif')
pae_pth = os.path.join(pred_dir, f'pae_{name}_model_{i}.npz')
pde_pth = os.path.join(pred_dir, f'pde_{name}_model_{i}.npz')
plddt_pth = os.path.join(pred_dir, f'plddt_{name}_model_{i}.npz')
name_rank_map[new_name].append({'confidence': cnf_pth,
'affinity' : aff_pth,
'cif_model' : mdl_pth,
'pae_pth' : pae_pth,
'pde_pth' : pde_pth,
'plddt_pth' : plddt_pth})
return name_rank_map
def update_output_name_dropdown(read_vhts: bool):
name_rank_f_map = read_output_files(read_vhts)
return (gr.update(choices=list(name_rank_f_map)),
gr.update(choices=['Rank 1'], value='Rank 1'),
name_rank_f_map)
def update_name_rank_dropdown(name: str, name_rank_f_map: dict):
total_rank = len(name_rank_f_map[name])
return gr.update(choices=[f'Rank {i}' for i in range(1, total_rank + 1)])
def update_result_visualization(name: str, rank_name: str, name_rank_f_map: dict):
if not rank_name.strip():
return [gr.update()] * 8
rank = int(rank_name.split(' ')[-1]) - 1
conf_metrics = name_rank_f_map[name][rank]
if rank+1 > len(conf_metrics):
return [gr.update()] * 8
with open(conf_metrics['confidence']) as f:
conf_dict = json.load(f)
overall_conf, chain_conf, pair_chain_conf = [], [], []
for conf_id, value in conf_dict.items():
if isinstance(value, float):
overall_conf.append([conf_id, f'{value:.4f}'])
elif conf_id == 'chains_ptm':
for chain_n, ptm_value in value.items():
chain_conf.append([f'{int(chain_n)+1}', f'{ptm_value:.4f}'])
elif conf_id == 'pair_chains_iptm':
for i, all_ptm_value in enumerate(value.values()):
pair_chain_conf.append([])
for single_ptm_value in all_ptm_value.values():
pair_chain_conf[i].append(f'{single_ptm_value:.4f}')
aff_f = conf_metrics['affinity']
if aff_f is not None:
aff_update = []
with open(aff_f) as f:
aff_data = json.load(f)
for aff_metric, aff_value in aff_data.items():
aff_update.append([aff_metric, f'{aff_value:.4f}'])
# combined_score = max((-aff_data['affinity_pred_value']-2)/4, 0) * aff_data['affinity_probability_binary']
# aff_update.append(['Overall Score', f'{combined_score:.4f}'])
aff_update = gr.update(value=aff_update, visible=True)
else:
aff_update = gr.update(visible=False)
with open(conf_metrics['cif_model']) as f:
mdl_strs = f.read()
cif_base64 = base64.b64encode(mdl_strs.encode()).decode('utf-8')
yield (get_molstar_html(cif_base64), gr.update(''), gr.update(''),
gr.update(''), gr.update(''), gr.update(''), gr.update(''), gr.update(''))
length_split = [0]
chain_entity_map = {}
last_res, last_c, i = None, None, 0
for line in mdl_strs.split('\n'):
if line.startswith(('ATOM', 'HETATM')):
if line.strip() == '#':
break
all_splitted = line.strip().split(' ')
res_id, entity_id, c = all_splitted[6], all_splitted[14], all_splitted[15]
chain_entity_map[c] = entity_id
if last_c is not None and last_c != c:
length_split.append(int(last_res) if last_res != '.' else i)
i = 0
last_c = c
last_res = res_id
if res_id == '.':
i += 1
elif line == '_atom_type.symbol':
if last_c is not None:
length_split.append(int(last_res) if last_res != '.' else i)
break
length_split = np.cumsum(length_split)
pae_mat = np.load(conf_metrics['pae_pth'])['pae']
pde_mat = np.load(conf_metrics['pde_pth'])['pde']
total_length = pae_mat.shape[0]
plddt_array = np.load(conf_metrics['plddt_pth'])['plddt']
pae_fig = px.imshow(pae_mat, color_continuous_scale='Greens_r',
range_color=[0.25, 31.75], labels={'color': 'PAE (Γ
)'})
for i in range(len(length_split)-2):
end = length_split[i+1]
pae_fig.add_shape(type='line', x0=0, y0=end-0.5,
x1=total_length-0.5, y1=end-0.5,
line=dict(color="Cornflowerblue"))
pae_fig.add_shape(type='line', x0=end-0.5, y0=0,
x1=end-0.5, y1=total_length-0.5,
line=dict(color="Cornflowerblue"))
pde_fig = px.imshow(pde_mat, color_continuous_scale='Greens_r',
range_color=[0.25, 31.75], labels={'color': 'PDE (Γ
)'})
for i in range(len(length_split)-2):
end = length_split[i+1]
pde_fig.add_shape(type='line', x0=0, y0=end-0.5,
x1=total_length-0.5, y1=end-0.5,
line=dict(color="Cornflowerblue"))
pde_fig.add_shape(type='line', x0=end-0.5, y0=0,
x1=end-0.5, y1=total_length-0.5,
line=dict(color="Cornflowerblue"))
plddt_fig = go.Figure()
all_chains = list(chain_entity_map)
for i in range(len(length_split)-1):
curr_c = all_chains[i]
splitted_plddt = plddt_array[length_split[i]:length_split[i+1]]
x_vals = np.arange(length_split[i]+1, length_split[i+1]+1)
mode = 'lines' if splitted_plddt.shape[0] > 1 else 'markers'
plddt_fig.add_trace(go.Scatter(x=x_vals,
y=splitted_plddt,
mode=mode,
name=f'Chain {curr_c} (Entity {chain_entity_map[curr_c]})'))
pae_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0))
pde_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0))
plddt_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0),
xaxis=dict(title=dict(text='Residue')),
yaxis=dict(title=dict(text='pLDDT')),
template='simple_white')
yield (gr.update(), overall_conf, chain_conf,
gr.DataFrame(value=pair_chain_conf,
headers=[f'{i+1}' for i in range(len(chain_conf))],
show_row_numbers=True, column_widths=['30px'] * len(chain_conf)),
aff_update, pae_fig, pde_fig, plddt_fig)
### vHTS Processing ###
def read_vhts_directory():
vhts_name_df = {}
vhts_name_pth_map = {}
for out_f_or_d in os.listdir(output_dir):
rng_dir = os.path.join(output_dir, out_f_or_d)
# If 'vHTS' is at the end of user-defined complex name then this will match too.
# Make a new dir specifically for vHTS in the future?
is_vhts = out_f_or_d.rsplit('_')[-2] == 'vHTS'
if os.path.isdir(rng_dir) and is_vhts:
for target_pth in os.listdir(rng_dir):
if 'boltz_results_' not in target_pth:
continue
name = target_pth.split('boltz_results_', 1)[-1].rsplit('_', 2)[0]
if name in vhts_name_df:
i = 2
new_name = f'{name}_{i}'
while new_name in vhts_name_df:
i += 1
new_name = f'{name}_{i}'
name = new_name
pred_dir = os.path.join(rng_dir, target_pth, 'predictions')
df_data = {'Name': [], 'ligand ipTM': [], 'binding prob.': [], 'binding aff.': []}
vhts_name_pth_map[name] = {'Name': [], 'conf': [], 'aff': [],
'struct': [], 'pae': [], 'pde': [], 'plddt': []}
for n in os.listdir(pred_dir):
docked_dir = Path(os.path.join(pred_dir, n))
if os.path.isdir(docked_dir):
conf_pth = docked_dir / f'confidence_{n}_model_0.json'
aff_pth = docked_dir / f'affinity_{n}.json'
struct_pth = docked_dir / f'{n}_model_0.cif'
pae_pth = docked_dir / f'pae_{n}_model_0.npz'
pde_pth = docked_dir / f'pde_{n}_model_0.npz'
plddt_pth = docked_dir / f'plddt_{n}_model_0.npz'
with open(conf_pth) as f:
ligand_iptm = json.load(f)['ligand_iptm']
with open(aff_pth) as f:
aff_data = json.load(f)
binding_aff = aff_data['affinity_pred_value']
binding_prob = aff_data['affinity_probability_binary']
for k, v in zip(df_data, [n, ligand_iptm, binding_prob, binding_aff]):
df_data[k].append(v)
vhts_name_pth_map[name][n] = {'conf' : conf_pth,
'aff' : aff_pth,
'struct': struct_pth,
'pae' : pae_pth,
'pde' : pde_pth,
'plddt' : plddt_pth}
df_data['Parent'] = [name] * len(df_data['Name'])
vhts_name_df[name] = pd.DataFrame(df_data)
return vhts_name_df, vhts_name_pth_map, gr.update(choices=list(vhts_name_df), value=None)
def update_vhts_df_with_selects(names: list[str], name_df_map: dict):
if not names:
return pd.DataFrame()
return pd.concat([name_df_map[n] for n in names]).reset_index(drop=True)
def update_vhts_result_visualization(name_fpth_map: dict, evt: gr.SelectData):
row_value = evt.row_value
if not row_value[0]:
return [gr.update()] * 9
parent, name = row_value[-1], row_value[0]
conf_metrics = name_fpth_map[parent][name]
with open(conf_metrics['conf']) as f:
conf_dict = json.load(f)
overall_conf, chain_conf, pair_chain_conf = [], [], []
for conf_id, value in conf_dict.items():
if isinstance(value, float):
overall_conf.append([conf_id, f'{value:.4f}'])
elif conf_id == 'chains_ptm':
for chain_n, ptm_value in value.items():
chain_conf.append([f'{int(chain_n)+1}', f'{ptm_value:.4f}'])
elif conf_id == 'pair_chains_iptm':
for i, all_ptm_value in enumerate(value.values()):
pair_chain_conf.append([])
for single_ptm_value in all_ptm_value.values():
pair_chain_conf[i].append(f'{single_ptm_value:.4f}')
aff_f = conf_metrics['aff']
aff_update = []
with open(aff_f) as f:
aff_data = json.load(f)
for aff_metric, aff_value in aff_data.items():
aff_update.append([aff_metric, f'{aff_value:.4f}'])
aff_update = gr.update(value=aff_update, visible=True)
with open(conf_metrics['struct']) as f:
mdl_strs = f.read()
cif_base64 = base64.b64encode(mdl_strs.encode()).decode('utf-8')
length_split = [0]
chain_entity_map = {}
last_res, last_c, i = None, None, 0
for line in mdl_strs.split('\n'):
if line.startswith(('ATOM', 'HETATM')):
if line.strip() == '#':
break
all_splitted = line.strip().split(' ')
res_id, entity_id, c = all_splitted[6], all_splitted[14], all_splitted[15]
chain_entity_map[c] = entity_id
if last_c is not None and last_c != c:
length_split.append(int(last_res) if last_res != '.' else i)
i = 0
last_c = c
last_res = res_id
if res_id == '.':
i += 1
elif line == '_atom_type.symbol':
if last_c is not None:
length_split.append(int(last_res) if last_res != '.' else i)
break
length_split = np.cumsum(length_split)
pae_mat = np.load(conf_metrics['pae'])['pae']
pde_mat = np.load(conf_metrics['pde'])['pde']
total_length = pae_mat.shape[0]
plddt_array = np.load(conf_metrics['plddt'])['plddt']
pae_fig = px.imshow(pae_mat, color_continuous_scale='Greens_r',
range_color=[0.25, 31.75], labels={'color': 'PAE (Γ
)'})
for i in range(len(length_split)-2):
end = length_split[i+1]
pae_fig.add_shape(type='line', x0=0, y0=end-0.5,
x1=total_length-0.5, y1=end-0.5,
line=dict(color="Cornflowerblue"))
pae_fig.add_shape(type='line', x0=end-0.5, y0=0,
x1=end-0.5, y1=total_length-0.5,
line=dict(color="Cornflowerblue"))
pde_fig = px.imshow(pde_mat, color_continuous_scale='Greens_r',
range_color=[0.25, 31.75], labels={'color': 'PDE (Γ
)'})
for i in range(len(length_split)-2):
end = length_split[i+1]
pde_fig.add_shape(type='line', x0=0, y0=end-0.5,
x1=total_length-0.5, y1=end-0.5,
line=dict(color="Cornflowerblue"))
pde_fig.add_shape(type='line', x0=end-0.5, y0=0,
x1=end-0.5, y1=total_length-0.5,
line=dict(color="Cornflowerblue"))
plddt_fig = go.Figure()
all_chains = list(chain_entity_map)
for i in range(len(length_split)-1):
curr_c = all_chains[i]
splitted_plddt = plddt_array[length_split[i]:length_split[i+1]]
x_vals = np.arange(length_split[i]+1, length_split[i+1]+1)
mode = 'lines' if splitted_plddt.shape[0] > 1 else 'markers'
plddt_fig.add_trace(go.Scatter(x=x_vals,
y=splitted_plddt,
mode=mode,
name=f'Chain {curr_c} (Entity {chain_entity_map[curr_c]})'))
pae_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0))
pde_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0))
plddt_fig.update_layout(margin=dict(l=0, r=0, t=0, b=0),
xaxis=dict(title=dict(text='Residue')),
yaxis=dict(title=dict(text='pLDDT')),
template='simple_white')
return (get_molstar_html(cif_base64), overall_conf, chain_conf,
gr.DataFrame(value=pair_chain_conf,
headers=[f'{i+1}' for i in range(len(chain_conf))],
show_row_numbers=True, column_widths=['30px'] * len(chain_conf)),
aff_update, pae_fig, pde_fig, plddt_fig,
f'<span style="font-size:15px; font-weight:bold;">Visualization of {name}</span>')
def download_vhts_dataframe(inp_df: pd.DataFrame, format: str):
inp_df = pd.DataFrame(inp_df)
if format == 'CSV':
temp_dir = tempfile.gettempdir()
saved_pth = os.path.join(temp_dir, 'vHTS_result.csv')
inp_df.to_csv(saved_pth, index=False)
elif format == 'TSV':
temp_dir = tempfile.gettempdir()
saved_pth = os.path.join(temp_dir, 'vHTS_result.tsv')
inp_df.to_csv(saved_pth, index=False, sep='\t')
elif format == 'XLSX':
temp_dir = tempfile.gettempdir()
saved_pth = os.path.join(temp_dir, 'vHTS_result.xlsx')
inp_df.to_excel(saved_pth, index=False)
return gr.update(saved_pth, interactive=True)
### Download Output ###
def zip_selected_files(all_files_and_dirs: list, zipname_pth_map: dict):
rng_name = uuid.uuid4().hex[:8]
zipped_file = os.path.join(curr_dir, f'output_{rng_name}.zip')
final_files = []
for f_or_d in all_files_and_dirs:
if os.path.isfile(f_or_d):
final_files.append(f_or_d)
max_f_cnt_len = len(str(len(final_files)))
yield f'{0:{max_f_cnt_len}}/{len(final_files)}', gr.update(), gr.update()
c = 0
with zipfile.ZipFile(zipped_file, 'w', zipfile.ZIP_DEFLATED, compresslevel=9) as zip_f:
for file in final_files:
zip_f.write(file, os.path.relpath(file, output_dir))
c += 1
yield f'{c:{max_f_cnt_len}}/{len(final_files)}', gr.update(), gr.update()
zipname_pth_map[os.path.basename(zipped_file)] = zipped_file
yield 'Zipping done', zipped_file, zipname_pth_map
def zip_selected_option_files(names: list, name_pth_map: dict, zipname_pth_map: dict):
rng_name = uuid.uuid4().hex[:8]
zipped_file = os.path.join(curr_dir, f'output_{rng_name}.zip')
final_files = []
for n in names:
pred_dir = name_pth_map[n]
for root, _, files in os.walk(pred_dir):
for file in files:
final_files.append(os.path.join(root, file))
max_f_cnt_len = len(str(len(final_files)))
yield f'{0:{max_f_cnt_len}}/{len(final_files)}', gr.update(), gr.update()
c = 0
with zipfile.ZipFile(zipped_file, 'w', zipfile.ZIP_DEFLATED, compresslevel=9) as zip_f:
for file in final_files:
zip_f.write(file, os.path.relpath(file, output_dir))
c += 1
yield f'{c:{max_f_cnt_len}}/{len(final_files)}', gr.update(), gr.update()
zipname_pth_map[os.path.basename(zipped_file)] = zipped_file
yield 'Zipping done', zipped_file, zipname_pth_map
def remove_zip_file(gr_tmp_pth: str, zipname_pth_map: dict):
# Remove the zip file to save disk space since gr.File already copy it to a new temp location
basename = os.path.basename(gr_tmp_pth)
os.remove(zipname_pth_map[basename])
del zipname_pth_map[basename]
return zipname_pth_map
def _extract_pred_dirs():
name_path_map = {}
for out_f_or_d in os.listdir(output_dir):
rng_dir = os.path.join(output_dir, out_f_or_d)
is_vhts = out_f_or_d.rsplit('_')[-2] == 'vHTS'
if os.path.isdir(rng_dir):
for target_pth in os.listdir(rng_dir):
if 'boltz_results_' not in target_pth:
continue
if not is_vhts:
target_dir = os.path.join(rng_dir, target_pth)
pred_parent_dir = os.path.join(target_dir, 'predictions')
if not os.path.isdir(pred_parent_dir):
shutil.rmtree(target_dir)
continue
for name in os.listdir(pred_parent_dir):
pred_dir = os.path.join(pred_parent_dir, name)
if not os.path.isdir(pred_dir):
continue
if name in name_path_map:
i = 2
new_name = f'{name}_{i}'
while new_name in name_path_map:
i += 1
new_name = f'{name}_{i}'
else:
new_name = name
name_path_map[new_name] = pred_dir
else:
target_dir = os.path.join(rng_dir, target_pth)
pred_parent_dir = os.path.join(target_dir, 'predictions')
name = target_pth.split('boltz_results_', 1)[-1].rsplit('_', 1)[0]
if name in name_path_map:
i = 2
new_name = f'{name}_{i}'
while new_name in name_path_map:
i += 1
new_name = f'{name}_{i}'
name = new_name
name_path_map[name] = pred_parent_dir
return name_path_map
def update_file_tree_and_dropdown():
file_explorer = gr.FileExplorer(root_dir=output_dir,
label='Output Files',
interactive=True)
name_path_map = _extract_pred_dirs()
return file_explorer, gr.update(choices=list(name_path_map)), name_path_map
### Utilities ###
def rdkit_embed_molecule(lig):
try:
report = AllChem.EmbedMolecule(lig, useRandomCoords=True)
if report == -1:
return None
else:
return lig
except Exception as e:
return None
def rdkit_embed_with_timeout(lig, timeout):
with ThreadPoolExecutor(max_workers=1) as executor:
future = executor.submit(rdkit_embed_molecule, lig)
try:
result = future.result(timeout=timeout)
return result
except Exception as e:
future.cancel()
return None
def reverse_complementary_nucleic_acid(inp_na: str, type: str):
if not inp_na.strip():
return ''
inp_na = inp_na.strip().upper()
for i, c in enumerate(inp_na):
if c not in 'ACTGU':
return f'Invalid nucleic acid sequence! Position {i+1} is "{c}".'
if type == 'Match Input':
if 'U' in inp_na and 'T' in inp_na:
return ('Both "U" and "T" are presented in input sequence!\n'
'Please manually specify which type of nucleic acid is required.')
elif 'U' in inp_na:
type = 'RNA'
else:
type = 'DNA'
mapping_dict = rev_comp_map[type]
return ''.join(mapping_dict[c] for c in inp_na[::-1])
def get_ligand_molstar_html(ccd_id: str):
return f"""
<iframe
id="molstar_frame"
style="width: 100%; height: 400px; border: none;"
srcdoc='
<!DOCTYPE html>
<html>
<head>
<script src="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.js"></script>
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.css">
</head>
<body>
<div id="Viewer" style="width: 300px; height: 300px; position: center"></div>
<script>
(async function() {{
const viewer = new rcsbMolstar.LigandViewer("Viewer",
{{showWelcomeToast: false,
layoutShowControls: false}});
const ccdID = "{ccd_id}";
try {{
await viewer.loadLigandId(ccdID);
}} catch (error) {{
console.error("Error loading structure:", error);
}}
}})();
</script>
</body>
</html>
'>
</iframe>"""
def get_mol_molstar_html(mol_str: str):
mol_js_string = json.dumps(mol_str)
return f"""
<iframe
style="width: 100%; height: 400px; border: none;"
srcdoc='
<!DOCTYPE html>
<html>
<head>
<script src="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.js"></script>
<link rel="stylesheet" href="https://cdn.jsdelivr.net/npm/@rcsb/rcsb-molstar/build/dist/viewer/rcsb-molstar.css">
</head>
<body>
<div id="Viewer" style="width: 100%; height: 100%;"></div>
<script>
(async function() {{
const viewer = new rcsbMolstar.Viewer("Viewer", {{
showWelcomeToast: false,
layoutShowControls: false
}});
try {{
await viewer.loadStructureFromData({mol_js_string}, "mol", false);
viewer.plugin.managers.interactivity.setProps({{ granularity: "element" }});
}} catch (err) {{
console.error("Mol* load error:", err);
}}
}})();
</script>
</body>
</html>
'>
</iframe>
"""
def draw_ccd_mol_3d(ccd_id: str):
ccd_id = ccd_id.upper()
yield get_ligand_molstar_html(ccd_id), pd.DataFrame()
cif_url = f'https://files.rcsb.org/ligands/download/{ccd_id}.cif'
result = requests.get(cif_url)
if result.status_code == 404:
yield get_ligand_molstar_html(''), pd.DataFrame()
data = cif.read_string(result.text)[ccd_id]
chem_descriptor_prefix = '_pdbx_chem_comp_descriptor'
looped_name = ['type', 'program', 'descriptor']
data_dict = {}
for name in looped_name:
loop = data.find_values(f'{chem_descriptor_prefix}.{name}')
data_dict[name.capitalize()] = [i.replace('"', '') for i in list(loop)]
yield gr.update(), pd.DataFrame(data_dict)
def draw_smiles_3d(smiles_str: str):
mol = Chem.MolFromSmiles(smiles_str)
if mol is None:
yield get_mol_molstar_html(''), gr.update(value=pd.DataFrame({'Property': ['Error'],
'Value': ['Invalid Molecule!']}))
else:
mol = Chem.AddHs(mol)
data_dict = {'Property': list(property_functions), 'Value': []}
for func in property_functions.values():
v = func(mol)
if isinstance(v, float):
v = round(v, 4)
data_dict['Value'].append(v)
yield get_mol_molstar_html(''), gr.update(value=pd.DataFrame(data_dict))
new_mol = rdkit_embed_with_timeout(mol, 60)
if new_mol is None:
mol = Chem.RemoveHs(mol) # If embedding failed / timeout, just use 2D coord
else:
mol = Chem.RemoveHs(new_mol)
mol_str = Chem.MolToMolBlock(mol)
yield get_mol_molstar_html(mol_str), gr.update()
### Boltz Interface ###
with gr.Blocks(css=css, theme=gr.themes.Default()) as Interface:
gr.Markdown('<span style="font-size:25px; font-weight:bold;">Boltz Interface</span>')
with gr.Tab('Single Complex'):
gr.Markdown('<span style="font-size:20px; font-weight:bold;">Basic Settings</span>')
template_name_chain_dict, template_name_path_dict,\
template_name_usage_dict, template_name_setting_dict = \
gr.State({}), gr.State({}), gr.State({}), gr.State({})
chain_res_dict = gr.State({})
with gr.Accordion('Template', open=False):
with gr.Row():
with gr.Group():
template_file = gr.Files(label='mmCIF tempalte(s)', file_types=['.cif'],
interactive=True)
template_dropdown = gr.Dropdown(label='Template Name', interactive=True)
with gr.Group():
use_template_chekcbox = gr.Checkbox(False, label='Use template',
interactive=False)
target_chain_ids = gr.Dropdown(label='Target Chain IDs',
multiselect=True, interactive=True)
template_chain_ids = gr.Dropdown(label='Template Chain IDs',
multiselect=True, interactive=True)
with gr.Accordion('Constraints', open=False):
with gr.Row(equal_height=True):
with gr.Column(scale=1):
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Bond conditioning</span>')
with gr.Group():
with gr.Row():
with gr.Column(min_width=60):
atom1_chain_dropdown = gr.Dropdown(label='Atom1 Chain',
interactive=True)
atom1_res_dropdown = gr.Dropdown(label='Atom1 Residue',
interactive=True)
atom1_atmname_text = gr.Text(label='Atom1 Name',
interactive=True)
with gr.Column(min_width=60):
atom2_chain_dropdown = gr.Dropdown(label='Atom2 Chain',
interactive=True)
atom2_res_dropdown = gr.Dropdown(label='Atom2 Residue',
interactive=True)
atom2_atmname_text = gr.Text(label='Atom2 Name',
interactive=True)
atom1_chain_dropdown.change(update_bond_sequence_length_with_chain,
inputs=[atom1_chain_dropdown, chain_res_dict],
outputs=atom1_res_dropdown)
atom2_chain_dropdown.change(update_bond_sequence_length_with_chain,
inputs=[atom2_chain_dropdown, chain_res_dict],
outputs=atom2_res_dropdown)
with gr.Column(scale=1):
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Pocket conditioning</span>')
pocket_binder = gr.Dropdown(label='Binder',
interactive=True)
pocket_text = gr.Text(label='Target Pockets',
placeholder='B:12,B:23',
interactive=True)
pocket_max_distance = gr.Number(6, label='Max Distance (Γ
)',
interactive=True, minimum=1)
with gr.Column(scale=2):
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Contact Conditioning</span>')
with gr.Group():
with gr.Row():
contact_1_dropdown = gr.Dropdown(label='Chain 1',
interactive=True)
contact_1_text = gr.Text(label='Reside IDX/Atom Name')
with gr.Row():
contact_2_dropdown = gr.Dropdown(label='Chain 2',
interactive=True)
contact_2_text = gr.Text(label='Reside IDX/Atom Name')
contact_max_distance = gr.Number(6, label='Max Distance (Γ
)',
interactive=True, minimum=1)
with gr.Row():
with gr.Column():
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Name, Affinity & Entities</span>')
single_complex_name = gr.Text(label='Name',
placeholder='Complex_1',
interactive=True)
affinity_binder = gr.Dropdown(label='Affinity Prediction Chain',
interactive=True)
mod_entity_number = gr.Number(1, label='Total Entity',
interactive=True, minimum=1, step=1)
def update_all_chains_dropdown(*all_entity_chain_values):
all_chains = set()
affinity_chains = set()
for i in range(0, len(all_entity_chain_values), 2):
entity, chain = all_entity_chain_values[i:i+2]
chains = [c.strip() for c in chain.split(',') if c.strip()]
all_chains.update(chains)
if entity in ['Ligand', 'CCD']:
affinity_chains.update(chains)
final_choices = [''] + sorted(all_chains)
aff_final = [''] + sorted(affinity_chains)
return (gr.update(choices=final_choices), gr.update(choices=aff_final),
gr.update(choices=final_choices), gr.update(choices=final_choices),
gr.update(choices=final_choices),)
@gr.render(inputs=mod_entity_number)
def append_new_entity(counts: int):
component_refs = []
for i in range(counts):
gr.Markdown(f'<span style="font-size:15px; font-weight:bold;">Entity {i+1}</span>', key=f'MK_{i}')
with gr.Row(key=f'Entity_{i}'):
with gr.Column(key=f'Entity_{i}_sub1', scale=1):
entity_menu = gr.Dropdown(entity_types,
label='Entity',
value=entity_types[0],
interactive=True,
key=f'ET_{i}', scale=1)
chain_name_text = gr.Text('',
label='Chains',
info='Press Enter to update Chains',
placeholder='A,B,C',
interactive=True,
key=f'Chain_{i}',
scale=1)
with gr.Column(key=f'Entity_{i}_sub2', scale=5):
with gr.Group(key=f'Entity_{i}_sub2_Group'):
sequence_text = gr.TextArea(label='Sequence',
placeholder='Input',
interactive=True,
lines=3,
key=f'SQ_{i}',
elem_classes='sequence')
highlight_text = gr.HighlightedText([('Input required!', 'X')],
label='Validation',
color_map={'β': 'green',
'X': 'red'},
key=f'HL_{i}',
elem_classes='validation',
show_legend=True)
with gr.Column(key=f'Entity_{i}_sub3', scale=1):
with gr.Group(key=f'Entity_{i}_sub3_group1'):
cyclic_ckbox = gr.Checkbox(False, label='Cyclic')
modification_text = gr.Text(label='Modifications (Residue:CCD)',
placeholder='2:ALY,15:MSE')
msa_file = gr.File(label='MSA File', file_types=['.a3m', '.csv'], height=92,
elem_classes='small-upload-style')
component_refs.extend([entity_menu, chain_name_text, sequence_text,
cyclic_ckbox, modification_text, msa_file])
entity_menu.change(change_sequence_label,
inputs=[entity_menu, sequence_text, cyclic_ckbox],
outputs=[sequence_text, highlight_text, cyclic_ckbox])
sequence_text.change(validate_sequence,
inputs=[entity_menu, sequence_text],
outputs=[highlight_text])
chain_name_text.submit(update_chain_seq_dict,
inputs=[entity_menu, chain_name_text,
sequence_text, chain_res_dict],
outputs=[chain_res_dict, atom1_chain_dropdown, atom2_chain_dropdown])
chain_name_text.input(update_chain_seq_dict,
inputs=[entity_menu, chain_name_text,
sequence_text, chain_res_dict],
outputs=[chain_res_dict, atom1_chain_dropdown, atom2_chain_dropdown])
entity_menu.change(update_chain_seq_dict,
inputs=[entity_menu, chain_name_text,
sequence_text, chain_res_dict],
outputs=[chain_res_dict, atom1_chain_dropdown, atom2_chain_dropdown])
gr.HTML("<hr>")
chain_components = [comp for i, comp in enumerate(component_refs) if i % 6 <= 1]
entity_components = [comp for i, comp in enumerate(component_refs) if i % 6 == 0]
for i in range(0, len(chain_components), 2):
chain_input = chain_components[i+1]
entity_menu = entity_components[i//2]
chain_input.submit(update_all_chains_dropdown,
inputs=chain_components,
outputs=[pocket_binder, affinity_binder,
contact_1_dropdown, contact_2_dropdown,
target_chain_ids])
chain_input.input(update_all_chains_dropdown,
inputs=chain_components,
outputs=[pocket_binder, affinity_binder,
contact_1_dropdown, contact_2_dropdown,
target_chain_ids])
entity_menu.change(update_all_chains_dropdown,
inputs=chain_components,
outputs=[pocket_binder, affinity_binder,
contact_1_dropdown, contact_2_dropdown,
target_chain_ids])
def write_yaml_func(binder, target, pocket_max_d, aff_binder,
cont_1_c, cont_1_r, cont_2_c, cont_2_r, contact_max_dist,
template_name_path_dict, template_name_usage_dict,
template_name_setting_dict,
bond_atom1_chain, bond_atom1_res, bond_atom1_name,
bond_atom2_chain, bond_atom2_res, bond_atom2_name,
*all_components):
all_components = list(all_components)
# TODO: Add more advanced format validation functions!
# constraints --> pocket
if binder and target:
contacts = []
for c_res in target.split(','):
if ':' not in c_res:
return ('Invalid target pocket, please use ":" to '
'separate target chain and target residue!\n'
'E.g., B:12,C:13')
c, r = c_res.split(':')
contacts.append([c, int(r)])
data_dict = {'sequences': [],
'constraints': [{'pocket': {'binder' : binder,
'contacts' : contacts,
'max_distance': pocket_max_d}}]}
else:
data_dict = {'sequences': []}
# constraints --> contact
if cont_1_c and cont_1_r.strip() and cont_2_c and cont_2_r.strip():
cont_1_r = cont_1_r.strip()
cont_2_r = cont_2_r.strip()
if cont_1_r.isdigit():
cont_1_r = int(cont_1_r)
if cont_2_r.isdigit():
cont_2_r = int(cont_2_r)
contact_dict = {'contact': {'token1': [cont_1_c, cont_1_r],
'token2': [cont_2_c, cont_2_r],
'max_distance': contact_max_dist}}
if 'constraints' in data_dict:
data_dict['constraints'].append(contact_dict)
else:
data_dict['constraints'] = [contact_dict]
# constraints --> bond
if all((bond_atom1_chain, bond_atom1_res, bond_atom1_name,
bond_atom2_chain, bond_atom2_res, bond_atom2_name)):
bond_dict = {'bond': {'atom1': [bond_atom1_chain, bond_atom1_res, bond_atom1_name.strip()],
'atom2': [bond_atom2_chain, bond_atom2_res, bond_atom2_name.strip()],}}
if 'constraints' in data_dict:
data_dict['constraints'].append(bond_dict)
else:
data_dict['constraints'] = [bond_dict]
# properties
if aff_binder:
data_dict.update({'properties': [{'affinity': {'binder': aff_binder}}]})
# templates
all_templates = []
for name in template_name_path_dict:
if template_name_usage_dict[name]:
curr_template = {'cif': template_name_path_dict[name]}
chain_template_id_dict = template_name_setting_dict[name]
if chain_template_id_dict['chain_id']:
curr_template['chain_id'] = chain_template_id_dict['chain_id']
if chain_template_id_dict['template_id']:
curr_template['template_id'] = chain_template_id_dict['template_id']
all_templates.append(curr_template)
if all_templates:
data_dict.update({'templates': all_templates})
existing_chains = []
msa_rng_name = uuid.uuid4().hex[:8]
for i in range(0, len(all_components), 6):
entity, chain, seq, cyclic, mod, msa_pth = all_components[i:i+6]
seq = seq.strip()
# set entity type
if entity == 'CCD':
entity_type = 'ligand'
else:
entity_type = entity.lower()
# set chain id
chains = chain.split(',')
if len(chains) == 1:
id = chain.strip()
if id in existing_chains:
return f'Chain {id} of Entity {i//6+1} already existed!'
existing_chains.append(id)
else:
id = [c.strip() for c in chains]
for _i in id:
if id.count(_i) > 1:
return f'Duplicate chain found within Entity {i//6+1}!'
if _i in existing_chains:
return f'Chain {id} of Entity {i//6+1} already existed!'
existing_chains.extend(id)
# set key of sequence ('sequence', 'ccd' or 'smiles')
if not seq:
return f'Entity {i//5+1} is empty!'
if entity == 'CCD':
seq_key = 'ccd'
seq = seq.upper()
if not re.fullmatch(r'(?:[A-Z0-9]{3}|[A-Z0-9]{5})|[A-Z]{2}', seq):
return f'Entity {i//5+1} is not a valid CCD ID!'
elif entity == 'Ligand':
seq_key = 'smiles'
if Chem.MolFromSmiles(seq) is None:
return f'Entity {i//5+1} is not a valid SMILES!'
else:
seq = seq.upper()
seq_key = 'sequence'
valid_strs = allow_char_dict[entity]
for char in seq:
if char not in valid_strs:
return f'Entity {i//5+1} is not a valid {entity}!'
# set modification
if mod:
modifications = []
all_mods = mod.split(',')
for pos_ccd in all_mods:
if ':' not in pos_ccd:
return (f'Invalid modification for Entity {i//6+1}, please use ":" to '
f'separate residue and CCD!\n')
pos, ccd = pos_ccd.split(':')
modifications.append({'position': int(pos), 'ccd': ccd})
else:
modifications = None
if entity_type == 'ligand':
curr_dict = {entity_type: {'id' : id,
seq_key : seq,}}
else:
curr_dict = {entity_type: {'id' : id,
seq_key : seq.upper(),
'cyclic': cyclic}}
if msa_pth and entity_type == 'protein':
target_msa = os.path.join(msa_dir, msa_rng_name, os.path.basename(msa_pth))
os.makedirs(os.path.dirname(target_msa), exist_ok=True)
os.rename(msa_pth, target_msa)
curr_dict[entity_type]['msa'] = target_msa
if modifications is not None:
curr_dict[entity_type]['modifications'] = modifications
data_dict['sequences'].append(curr_dict)
yaml_string = safe_dump(data_dict, sort_keys=False, indent=4)
return yaml_string
write_yaml_button.click(write_yaml_func,
inputs=[pocket_binder, pocket_text,
pocket_max_distance, affinity_binder,
contact_1_dropdown, contact_1_text,
contact_2_dropdown, contact_2_text,
contact_max_distance,
template_name_path_dict,
template_name_usage_dict,
template_name_setting_dict,
atom1_chain_dropdown, atom1_res_dropdown, atom1_atmname_text,
atom2_chain_dropdown, atom2_res_dropdown, atom2_atmname_text,
*component_refs],
outputs=[yaml_text])
with gr.Row():
with gr.Column():
write_yaml_button = gr.Button('Write YAML')
add_single_to_bacth_button = gr.Button('Add to Batch')
run_single_boltz_button = gr.Button('Run Boltz', interactive=False)
yaml_text = gr.Code(label='YAML Output',
scale=4,
language='yaml',
interactive=True, max_lines=15)
single_boltz_log = gr.Textbox(label='Prediction Log', lines=10, max_lines=10,
autofocus=False, elem_classes='log')
template_file.upload(read_tempaltes,
inputs=[template_file,
template_name_chain_dict, template_name_path_dict,
template_name_usage_dict, template_name_setting_dict],
outputs=[template_dropdown,
template_name_chain_dict, template_name_path_dict,
template_name_usage_dict, use_template_chekcbox,
template_name_setting_dict])
use_template_chekcbox.input(update_template_chain_ids_and_settings,
inputs=[use_template_chekcbox, target_chain_ids, template_chain_ids,
template_dropdown, template_name_usage_dict, template_name_setting_dict],
outputs=[template_name_usage_dict, template_name_setting_dict])
target_chain_ids.input(update_template_chain_ids_and_settings,
inputs=[use_template_chekcbox, target_chain_ids, template_chain_ids,
template_dropdown, template_name_usage_dict, template_name_setting_dict],
outputs=[template_name_usage_dict, template_name_setting_dict])
template_chain_ids.input(update_template_chain_ids_and_settings,
inputs=[use_template_chekcbox, target_chain_ids, template_chain_ids,
template_dropdown, template_name_usage_dict, template_name_setting_dict],
outputs=[template_name_usage_dict, template_name_setting_dict])
template_dropdown.change(update_template_dropdown,
inputs=[template_dropdown, template_name_chain_dict,
template_name_usage_dict, template_name_setting_dict],
outputs=[use_template_chekcbox, target_chain_ids, template_chain_ids])
single_complex_name.input(check_yaml_strings,
inputs=[yaml_text, single_complex_name],
outputs=run_single_boltz_button)
yaml_text.change(check_yaml_strings,
inputs=[yaml_text, single_complex_name],
outputs=run_single_boltz_button)
run_single_boltz_button.click(execute_single_boltz,
inputs=[single_complex_name, yaml_text,
*all_boltz_parameters],
outputs=[run_single_boltz_button, single_boltz_log])
with gr.Tab('Batch Predict'):
batch_upload_files = gr.State({})
processed_inp_files = gr.State([])
with gr.Row():
with gr.Column(scale=1):
mod_batch_total_files = gr.Number(0, label='Total Files',
scale=1, interactive=True,
minimum=0, step=1)
clear_batch_button = gr.Button('Clear')
upload_yaml_files = gr.Files(file_types=['.yaml', '.yml'],
label='Upload YAML files',
interactive=True, scale=2)
upload_yaml_files.upload(upload_multi_files,
inputs=[upload_yaml_files, mod_batch_total_files],
outputs=[batch_upload_files, mod_batch_total_files, upload_yaml_files])
add_single_to_bacth_button.click(add_current_single_to_batch,
inputs=[single_complex_name, yaml_text,
batch_upload_files, mod_batch_total_files],
outputs=[batch_upload_files, mod_batch_total_files,
add_single_to_bacth_button])
clear_batch_button.click(clear_curr_batch_dict,
outputs=[batch_upload_files, mod_batch_total_files])
@gr.render(inputs=[batch_upload_files, mod_batch_total_files],
triggers=[clear_batch_button.click, mod_batch_total_files.change])
def create_new_batch_file_count(all_uploaded_files: dict, counts: int):
batched_files = []
total_uploaded = len(all_uploaded_files)
paired_all_files = list(all_uploaded_files.items())
pair_c = 0
for i in range(counts):
gr.Markdown(f'<span style="font-size:15px; font-weight:bold;">File {i+1}</span>', key=f'B_MK_{i}')
with gr.Row(key=f'B_File_{i}'):
if i >= counts - total_uploaded:
name_str = paired_all_files[min(pair_c, counts-1)][0]
yaml_str = paired_all_files[min(pair_c, counts-1)][1]
file_name_text = gr.Text(name_str,
label='Name',
interactive=True, scale=1,
key=f'name_{i}')
yaml_str_code = gr.Code(yaml_str,
label='YAML String',
language='yaml',
interactive=True, scale=4, max_lines=10,
key=f'yaml_{i}')
pair_c += 1
else:
file_name_text = gr.Text(label='Name',
interactive=True, scale=1,
key=f'name_{i}')
yaml_str_code = gr.Code(label='YAML String',
language='yaml',
interactive=True, scale=4, max_lines=10,
key=f'yaml_{i}')
batched_files.extend([file_name_text, yaml_str_code])
file_name_text.input(check_batch_yaml_and_name,
inputs=[yaml_str_code, file_name_text],
outputs=file_name_text)
yaml_str_code.input(check_batch_yaml_and_name,
inputs=[yaml_str_code, file_name_text],
outputs=file_name_text)
gr.HTML("<hr>")
def process_all_files(*all_components):
all_components = list(all_components)
final_result_map = {}
inp_rng_dir = os.path.join(input_dir, f"batch_{uuid.uuid4().hex[:8]}")
check_dir_exist_and_rename(inp_rng_dir)
for i in range(0, len(all_components), 2):
file_name, yaml_str = all_components[i:i+2]
file_name = file_name.strip()
yaml_valid = _check_yaml_strings(yaml_str)
if file_name and yaml_valid:
final_result_map[os.path.join(inp_rng_dir, f'{file_name}.yaml')] = yaml_str
for f_pth, yaml_content in final_result_map.items():
with open(f_pth, 'w') as f:
f.write(yaml_content)
return list(final_result_map), gr.update(interactive=True), list(final_result_map)
batch_process_all_files.click(process_all_files,
inputs=batched_files,
outputs=[batch_process_result,
batch_predict_all_files,
processed_inp_files])
with gr.Row():
with gr.Column(scale=1):
batch_process_all_files = gr.Button('Batch Process')
batch_predict_all_files = gr.Button('Batch Predict', interactive=False)
with gr.Column(scale=2):
batch_process_result = gr.File(label='Processed Files',
interactive=False, file_count='multiple')
multi_boltz_log = gr.Textbox(label='Prediction Log', lines=10, max_lines=10,
autofocus=False, elem_classes='log')
batch_predict_all_files.click(execute_multi_boltz,
inputs=[processed_inp_files,
*all_boltz_parameters],
outputs=[batch_predict_all_files, multi_boltz_log])
with gr.Tab('vHTS'):
gr.Markdown('<span style="font-size:20px; font-weight:bold;">Multiple molecules vs Single protein</span>')
with gr.Accordion('1. Ligand Settings', open=False):
with gr.Row():
with gr.Column():
with gr.Row():
ligand_file_type = gr.Dropdown(['Chemical files', 'Tabular files'],
'Chemical files',
label='Chemical format type',
interactive=True)
vhts_ligand_chain_text = gr.Text(label='Ligand Chain', placeholder='Z',
interactive=True, value='Z')
vhts_clear_ligand_df_btn = gr.Button('Clear Ligands')
ligand_dataframe = gr.DataFrame(headers=['Name', 'SMILES'],
max_height=300, interactive=True, min_width=250,
show_row_numbers=True)
with gr.Column():
chemical_file_upload_file = gr.File(label='Upload chemical file(s)', interactive=True,
file_count='multiple',
file_types=['.sdf', '.mol', '.smi', '.zip'])
with gr.Row():
tabular_chem_file_name_id = gr.Text(label='Column of Name',
interactive=True, visible=False)
tabular_chem_file_chem_id = gr.Text(label='Column of Chem String',
interactive=True, visible=False)
tabular_chem_file_delimiter = gr.Dropdown([',', '\t', ';', ' '], value=',',
label='Delimiter',
allow_custom_value=True, visible=False,
interactive=True)
ligand_file_type.input(update_chem_file_format, inputs=ligand_file_type,
outputs=[chemical_file_upload_file, tabular_chem_file_name_id,
tabular_chem_file_chem_id, tabular_chem_file_delimiter])
vhts_clear_ligand_df_btn.click(lambda x: pd.DataFrame(), inputs=ligand_dataframe, outputs=ligand_dataframe)
chemical_file_upload_file.upload(process_uploaded_ligand,
inputs=[chemical_file_upload_file, tabular_chem_file_name_id,
tabular_chem_file_chem_id, tabular_chem_file_delimiter,
ligand_dataframe],
outputs=[ligand_dataframe])
with gr.Accordion('2. Protein Settings', open=False):
with gr.Accordion('Template', open=False):
with gr.Row():
vhts_template_name_chain_dict, vhts_template_name_path_dict,\
vhts_template_name_usage_dict, vhts_template_name_setting_dict = \
gr.State({}), gr.State({}), gr.State({}), gr.State({})
with gr.Group():
vhts_template_file = gr.Files(label='mmCIF tempalte(s)', file_types=['.cif'],
interactive=True)
vhts_template_dropdown = gr.Dropdown(label='Template Name', interactive=True)
with gr.Group():
vhts_use_template_chekcbox = gr.Checkbox(False, label='Use template',
interactive=False)
vhts_target_chain_ids = gr.Dropdown(label='Target Chain IDs',
multiselect=True, interactive=True)
vhts_template_chain_ids = gr.Dropdown(label='Template Chain IDs',
multiselect=True, interactive=True)
with gr.Row():
with gr.Column(scale=1):
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Pocket Conditioning & Entity Count</span>')
vhts_pocket_text = gr.Text(label='Target Pockets',
placeholder='B:12,B:23',
interactive=True)
vhts_pocket_max_distance = gr.Number(6, label='Max Distance (Γ
)',
interactive=True, minimum=1)
vhts_entity_number = gr.Number(1, label='Total Entity',
interactive=True, minimum=1, step=1)
with gr.Column(scale=1):
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Contact Conditioning</span>')
with gr.Group():
with gr.Row():
vhts_contact_1_dropdown = gr.Dropdown(label='Chain 1',
interactive=True)
vhts_contact_1_text = gr.Text(label='Reside')
with gr.Row():
vhts_contact_2_dropdown = gr.Dropdown(label='Chain 2',
interactive=True)
vhts_contact_2_text = gr.Text(label='Reside')
vhts_contact_max_distance = gr.Number(6, label='Max Distance (Γ
)',
interactive=True, minimum=1)
def vhts_update_all_chains_dropdown(*all_entity_chain_values):
all_chains = set()
for chain in all_entity_chain_values:
chains = [c.strip() for c in chain.split(',') if c.strip()]
all_chains.update(chains)
final_choices = [''] + sorted(all_chains)
return (gr.update(choices=final_choices), gr.update(choices=final_choices),
gr.update(choices=final_choices))
@gr.render(inputs=vhts_entity_number)
def vhts_append_new_entity(counts: int):
component_refs = []
for i in range(counts):
gr.Markdown(f'<span style="font-size:15px; font-weight:bold;">Entity {i+1}</span>', key=f'MK_{i}')
with gr.Row(key=f'Entity_{i}'):
with gr.Column(key=f'Entity_{i}_sub1', scale=1):
entity_menu = gr.Dropdown(entity_types,
label='Entity',
value=entity_types[0],
interactive=True,
key=f'ET_{i}', scale=1)
chain_name_text = gr.Text('',
label='Chains',
info='Press Enter to update Binder',
placeholder='A,B,C',
interactive=True,
key=f'Chain_{i}',
scale=1)
with gr.Column(key=f'Entity_{i}_sub2', scale=4):
with gr.Group(key=f'Entity_{i}_sub2_Group'):
sequence_text = gr.TextArea(label='Sequence',
placeholder='Input',
interactive=True,
lines=3,
key=f'SQ_{i}',
elem_classes='sequence')
highlight_text = gr.HighlightedText([('Input required!', 'X')],
label='Validation',
color_map={'β': 'green',
'X': 'red'},
key=f'HL_{i}',
elem_classes='validation',
show_legend=True)
with gr.Column(key=f'Entity_{i}_sub3', scale=1):
cyclic_ckbox = gr.Checkbox(False, label='Cyclic')
modification_text = gr.Text(label='Modifications (Residue:CCD)',
placeholder='2:ALY,15:MSE')
component_refs.extend([entity_menu, chain_name_text, sequence_text,
cyclic_ckbox, modification_text])
entity_menu.change(change_sequence_label,
inputs=[entity_menu, sequence_text, cyclic_ckbox],
outputs=[sequence_text, highlight_text, cyclic_ckbox])
sequence_text.change(validate_sequence,
inputs=[entity_menu, sequence_text],
outputs=highlight_text)
gr.HTML("<hr>")
chain_components = [comp for i, comp in enumerate(component_refs) if i % 5 == 1]
entity_components = [comp for i, comp in enumerate(component_refs) if i % 5 == 0]
for i, chain_input in enumerate(chain_components):
chain_input.submit(vhts_update_all_chains_dropdown,
inputs=chain_components,
outputs=[vhts_contact_1_dropdown, vhts_contact_2_dropdown,
vhts_target_chain_ids])
entity_components[i].change(vhts_update_all_chains_dropdown,
inputs=chain_components,
outputs=[vhts_contact_1_dropdown, vhts_contact_2_dropdown,
vhts_target_chain_ids])
def write_yaml_func(binder, target, pocket_max_d, aff_binder,
cont_1_c, cont_1_r, cont_2_c, cont_2_r,
template_name_path_dict, template_name_usage_dict,
template_name_setting_dict,
*all_components):
all_components = list(all_components)
if not binder:
return 'Ligand chain must be provided!'
if binder and target:
contacts = []
for c_res in target.split(','):
if ':' not in c_res:
return ('Invalid target pocket, please use ":" to '
'separate target chain and target residue!\n'
'E.g., B:12,C:13')
c, r = c_res.split(':')
contacts.append([c, int(r)])
data_dict = {'sequences': [],
'constraints': [{'pocket': {'binder' : binder,
'contacts' : contacts,
'max_distance': pocket_max_d}}]}
else:
data_dict = {'sequences': []}
if aff_binder:
data_dict.update({'properties': [{'affinity': {'binder': aff_binder}}]})
if cont_1_c and cont_1_r.strip() and cont_2_c and cont_2_r.strip():
contact_dict = {'contact': {'token1': [cont_1_c, cont_1_r.strip()],
'token2': [cont_2_c, cont_2_r.strip()],}}
if 'constraints' in data_dict:
data_dict['constraints'].append(contact_dict)
else:
data_dict['constraints'] = [contact_dict]
all_templates = []
for name in template_name_path_dict:
if template_name_usage_dict[name]:
curr_template = {'cif': template_name_path_dict[name]}
chain_template_id_dict = template_name_setting_dict[name]
if chain_template_id_dict['chain_id']:
curr_template['chain_id'] = chain_template_id_dict['chain_id']
if chain_template_id_dict['template_id']:
curr_template['template_id'] = chain_template_id_dict['template_id']
all_templates.append(curr_template)
if all_templates:
data_dict.update({'templates': all_templates})
existing_chains = []
all_components += ['Ligand', binder, 'c1ccccc1', False, '']
for i in range(0, len(all_components), 5):
entity, chain, seq, cyclic, mod = all_components[i:i+5]
seq = seq.strip()
# set entity type
if entity == 'CCD':
entity_type = 'ligand'
else:
entity_type = entity.lower()
# set chain id
chains = chain.split(',')
if len(chains) == 1:
id = chain.strip()
if id in existing_chains:
return f'Chain {id} of Entity {i//5+1} already existed!'
existing_chains.append(id)
else:
id = [c.strip() for c in chains]
for _i in id:
if id.count(_i) > 1:
return f'Duplicate chain found within Entity {i//5+1}!'
if _i in existing_chains:
return f'Chain {id} of Entity {i//5+1} already existed!'
existing_chains.extend(id)
# set key of sequence ('sequence', 'ccd' or 'smiles')
if not seq:
return f'Entity {i//5+1} is empty!'
if entity == 'CCD':
seq = seq.upper()
seq_key = 'ccd'
if not re.fullmatch(r'(?:[A-Z0-9]{3}|[A-Z0-9]{5}|[A-Z]{2})', seq):
return f'Entity {i//5+1} is not a valid CCD ID!'
elif entity == 'Ligand':
seq_key = 'smiles'
if Chem.MolFromSmiles(seq) is None:
return f'Entity {i//5+1} is not a valid SMILES!'
else:
seq = seq.upper()
seq_key = 'sequence'
valid_strs = allow_char_dict[entity]
for char in seq:
if char not in valid_strs:
return f'Entity {i//5+1} is not a valid {entity}!'
# set modification
if mod:
modifications = []
all_mods = mod.split(',')
for pos_ccd in all_mods:
if ':' not in pos_ccd:
return (f'Invalid modification for Entity {i//5+1}, please use ":" to '
f'separate residue and CCD!\n')
pos, ccd = pos_ccd.split(':')
modifications.append({'position': int(pos), 'ccd': ccd})
else:
modifications = None
if entity_type == 'ligand':
curr_dict = {entity_type: {'id' : id,
seq_key : seq,}
}
else:
curr_dict = {entity_type: {'id' : id,
seq_key : seq.upper(),
'cyclic': cyclic}
}
if modifications is not None:
curr_dict[entity_type]['modifications'] = modifications
data_dict['sequences'].append(curr_dict)
yaml_string = safe_dump(data_dict, sort_keys=False, indent=4)
yaml_string = '#This is a demo file with the ligand replaced with benzene.\n' + yaml_string
return yaml_string
vhts_process_file_demo_button.click(write_yaml_func,
inputs=[vhts_ligand_chain_text, vhts_pocket_text,
vhts_pocket_max_distance, vhts_ligand_chain_text,
vhts_contact_1_dropdown, vhts_contact_1_text,
vhts_contact_2_dropdown, vhts_contact_2_text,
vhts_template_name_path_dict,
vhts_template_name_usage_dict,
vhts_template_name_setting_dict,
*component_refs],
outputs=vhts_yaml_demo_text)
with gr.Accordion('3. Process Settings & Start Screening', open=False):
with gr.Row():
with gr.Column(scale=1):
vhts_complex_prefix = gr.Text(label='Prefix',
info=('A prefix that will be added to the output directory '
'(quote not included)'),
placeholder='"Protein"_', interactive=True)
vhts_process_file_demo_button = gr.Button('Write Demo YAML')
vhts_start_predict_button = gr.Button('Run vHTS', interactive=False)
vhts_yaml_demo_text = gr.Code(label='Demo YAML output',
language='yaml',
scale=4, interactive=False, max_lines=15)
vhts_template_file.upload(read_tempaltes,
inputs=[vhts_template_file,
vhts_template_name_chain_dict, vhts_template_name_path_dict,
vhts_template_name_usage_dict, vhts_template_name_setting_dict],
outputs=[vhts_template_dropdown,
vhts_template_name_chain_dict, vhts_template_name_path_dict,
vhts_template_name_usage_dict, vhts_use_template_chekcbox,
vhts_template_name_setting_dict])
vhts_use_template_chekcbox.input(update_template_chain_ids_and_settings,
inputs=[vhts_use_template_chekcbox, vhts_target_chain_ids,
vhts_template_chain_ids,
vhts_template_dropdown, vhts_template_name_usage_dict,
vhts_template_name_setting_dict],
outputs=[vhts_template_name_usage_dict, vhts_template_name_setting_dict])
target_chain_ids.input(update_template_chain_ids_and_settings,
inputs=[vhts_use_template_chekcbox, vhts_target_chain_ids, vhts_template_chain_ids,
vhts_template_dropdown, vhts_template_name_usage_dict, vhts_template_name_setting_dict],
outputs=[vhts_template_name_usage_dict, vhts_template_name_setting_dict])
template_chain_ids.input(update_template_chain_ids_and_settings,
inputs=[vhts_use_template_chekcbox, vhts_target_chain_ids, vhts_template_chain_ids,
vhts_template_dropdown, vhts_template_name_usage_dict, vhts_template_name_setting_dict],
outputs=[vhts_template_name_usage_dict, vhts_template_name_setting_dict])
template_dropdown.change(update_template_dropdown,
inputs=[vhts_template_dropdown, vhts_template_name_chain_dict,
vhts_template_name_usage_dict, vhts_template_name_setting_dict],
outputs=[vhts_use_template_chekcbox, vhts_target_chain_ids, vhts_template_chain_ids])
vhts_boltz_log = gr.Textbox(label='Prediction Log', lines=10, max_lines=10,
autofocus=False, elem_classes='log')
ligand_dataframe.change(check_yaml_strings,
inputs=[vhts_yaml_demo_text, vhts_complex_prefix, ligand_dataframe],
outputs=vhts_start_predict_button)
vhts_complex_prefix.input(check_yaml_strings,
inputs=[vhts_yaml_demo_text, vhts_complex_prefix, ligand_dataframe],
outputs=vhts_start_predict_button)
vhts_yaml_demo_text.change(check_yaml_strings,
inputs=[vhts_yaml_demo_text, vhts_complex_prefix, ligand_dataframe],
outputs=vhts_start_predict_button)
vhts_start_predict_button.click(execute_vhts_boltz,
inputs=[vhts_complex_prefix, ligand_dataframe, vhts_ligand_chain_text,
vhts_yaml_demo_text, *all_boltz_parameters],
outputs=[vhts_start_predict_button, vhts_boltz_log])
with gr.Tab('Result Visualization'):
name_rank_f_map_state = gr.State({})
with gr.Row():
refresh_vis_button = gr.Button('Refresh', scale=1)
read_vhts_checkbox = gr.Checkbox(False, label='Read vHTS Result', interactive=True)
with gr.Column(scale=3):
...
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Select Name and Rank</span>')
with gr.Row():
result_name_dropdown = gr.Dropdown(label='Name',
info='Name of the complex in the output',
interactive=True)
result_rank_dropdown = gr.Dropdown(label='Rank',
info='Rank of the selected complex',
interactive=True)
gr.Markdown('<span style="font-size:15px; font-weight:bold;">Result</span>')
mol_star_html = gr.HTML(get_molstar_html(''))
with gr.Row():
conf_df = gr.DataFrame(headers=['Metric', 'Score'], label='Overall Metrics', scale=1)
with gr.Column(scale=2):
with gr.Row():
chain_metrics = gr.DataFrame(headers=['Chain Num.', 'pTM Score'],
label='Chain pTM', scale=1)
pair_chain_metrics = gr.DataFrame(headers=None,
label='Pairwise chain ipTM',
show_row_numbers=True, scale=2,
wrap=True)
aff_df = gr.DataFrame(headers=['Metric', 'Score'], label='Affinity Metrics')
with gr.Row():
pae_plot = gr.Plot(label='PAE', format='png')
pde_plot = gr.Plot(label='PDE', format='png')
plddt_plot = gr.Plot(label='pLDDT', format='png')
refresh_vis_button.click(update_output_name_dropdown,
inputs=read_vhts_checkbox,
outputs=[result_name_dropdown,
result_rank_dropdown,
name_rank_f_map_state])
result_name_dropdown.input(update_name_rank_dropdown,
inputs=[result_name_dropdown, name_rank_f_map_state],
outputs=result_rank_dropdown)
result_name_dropdown.input(update_result_visualization,
inputs=[result_name_dropdown, result_rank_dropdown, name_rank_f_map_state],
outputs=[mol_star_html, conf_df, chain_metrics, pair_chain_metrics,
aff_df, pae_plot, pde_plot, plddt_plot])
result_rank_dropdown.input(update_result_visualization,
inputs=[result_name_dropdown, result_rank_dropdown, name_rank_f_map_state],
outputs=[mol_star_html, conf_df, chain_metrics, pair_chain_metrics,
aff_df, pae_plot, pde_plot, plddt_plot])
with gr.Tab('vHTS Analysis'):
vhts_name_df_map, vhts_name_file_map = gr.State({}), gr.State({})
with gr.Row():
refresh_vhts_button = gr.Button('Refresh', scale=1)
with gr.Column(scale=3):
...
vhts_output_options = gr.Dropdown(label='vHTS Output', multiselect=True, interactive=True)
vhts_output_df = gr.DataFrame(label='vHTS Result Table', interactive=False,
headers=['Name', 'ligand ipTM',
'binding prob.',
'binding aff.', 'Parent'],
show_row_numbers=True, show_copy_button=True, show_search='filter')
# with gr.Row():
# vhts_table_download_format = gr.Dropdown(['', 'CSV', 'TSV', 'XLSX'], value='', label='Tabular Format')
# vhts_download_button = gr.DownloadButton('Download Tabular File', interactive=False)
# with gr.Column(scale=3):
# ...
vhts_header = gr.Markdown('<span style="font-size:15px; font-weight:bold;">Visualization</span>')
vhts_mol_star_html = gr.HTML(get_molstar_html(''))
with gr.Row():
vhts_conf_df = gr.DataFrame(headers=['Metric', 'Score'], label='Overall Metrics', scale=1)
with gr.Column(scale=2):
with gr.Row():
vhts_chain_metrics = gr.DataFrame(headers=['Chain Num.', 'pTM Score'],
label='Chain pTM', scale=1)
vhts_pair_chain_metrics = gr.DataFrame(headers=None,
label='Pairwise chain ipTM',
show_row_numbers=True, scale=2,
wrap=True)
vhts_aff_df = gr.DataFrame(headers=['Metric', 'Score'], label='Affinity Metrics')
with gr.Row():
vhts_pae_plot = gr.Plot(label='PAE', format='png')
vhts_pde_plot = gr.Plot(label='PDE', format='png')
vhts_plddt_plot = gr.Plot(label='pLDDT', format='png')
refresh_vhts_button.click(read_vhts_directory,
outputs=[vhts_name_df_map,
vhts_name_file_map,
vhts_output_options])
# vhts_table_download_format.input(download_vhts_dataframe,
# inputs=[vhts_output_df, vhts_table_download_format],
# outputs=vhts_download_button)
vhts_output_options.input(update_vhts_df_with_selects,
inputs=[vhts_output_options, vhts_name_df_map],
outputs=vhts_output_df)
vhts_output_df.select(update_vhts_result_visualization,
inputs=[vhts_name_file_map],
outputs=[vhts_mol_star_html, vhts_conf_df, vhts_chain_metrics,
vhts_pair_chain_metrics, vhts_aff_df, vhts_pae_plot,
vhts_pde_plot, vhts_plddt_plot, vhts_header])
with gr.Tab('Boltz Output'):
all_zipped_files_map = gr.State({})
with gr.Row():
refresh_button = gr.Button('Refresh', scale=1)
with gr.Column(scale=3):
...
with gr.Accordion('File List', open=False):
output_file_tree = gr.FileExplorer(root_dir=output_dir,
label='Output Files',
interactive=True)
with gr.Row():
with gr.Column(scale=1):
download_selected_button = gr.Button('Download')
zipping_progress = gr.Text(label='Zipping Progress', interactive=False)
download_zip_files = gr.File(label='Zipped File Download',
scale=3, file_count='single',
file_types=['.zip'], interactive=False)
with gr.Accordion('Directory List', open=False):
output_map = _extract_pred_dirs()
download_file_pth_map = gr.State(output_map)
output_directory_options = gr.Dropdown(choices=list(output_map),
label='Output Directories',
multiselect=True, interactive=True)
with gr.Row():
with gr.Column(scale=1):
download_selected_option_button = gr.Button('Download')
zipping_option_progress = gr.Text(label='Zipping Progress', interactive=False)
download_zip_option_files = gr.File(label='Zipped File Download',
scale=3, file_count='single',
file_types=['.zip'], interactive=False)
refresh_button.click(update_file_tree_and_dropdown,
outputs=[output_file_tree,
output_directory_options,
download_file_pth_map])
download_selected_button.click(zip_selected_files,
inputs=[output_file_tree, all_zipped_files_map],
outputs=[zipping_progress, download_zip_files, all_zipped_files_map])
download_selected_option_button.click(zip_selected_option_files,
inputs=[output_directory_options,
download_file_pth_map,
all_zipped_files_map],
outputs=[zipping_option_progress,
download_zip_option_files,
all_zipped_files_map])
download_zip_option_files.change(remove_zip_file,
inputs=[download_zip_option_files, all_zipped_files_map],
outputs=[all_zipped_files_map])
with gr.Tab('Boltz Paramters'):
with gr.Row():
with gr.Column():
gr.Markdown('<span style="font-size:20px; font-weight:bold;">System setting</span>')
device_number.render()
accelerator_type.render()
download_model_weight = gr.Button('Download Weight (Boltz-2)')
with gr.Column():
gr.Markdown('<span style="font-size:20px; font-weight:bold;">Boltz Parameters</span>')
boltz_method.render()
recycling_steps.render()
sampling_steps.render()
diffusion_samples.render()
step_scale.render()
num_workers.render()
preprocessing_threads.render()
affinity_mw_correction.render()
sampling_steps_affinity.render()
diffusion_samples_affinity.render()
no_trifast.render()
override.render()
use_potentials.render()
download_model_weight.click(manual_download_boltz_weights, outputs=download_model_weight)
with gr.Tab('Utilities'):
with gr.Accordion('Inverse Complement Nucleic Acid', open=False):
inp_nucleic_acid = gr.TextArea(label='Input DNA/RNA', lines=3, interactive=True)
with gr.Row(equal_height=True):
rev_comp_na_type = gr.Dropdown(['Match Input', 'DNA', 'RNA'], value='Match Input',
interactive=True, label='Nucleic Acid Type', scale=1)
rev_comp_na_text = gr.TextArea(label='Inverse Complement', lines=3,
show_copy_button=True, scale=5)
inp_nucleic_acid.input(reverse_complementary_nucleic_acid,
inputs=[inp_nucleic_acid, rev_comp_na_type],
outputs=rev_comp_na_text)
rev_comp_na_type.input(reverse_complementary_nucleic_acid,
inputs=[inp_nucleic_acid, rev_comp_na_type],
outputs=rev_comp_na_text)
with gr.Accordion('Display Tabular File', open=False):
with gr.Row():
utility_tabular_file = gr.File(label='Tabular file', interactive=True,
file_types=['.csv', '.tsv', '.txt'])
with gr.Column():
utility_custom_delimiter_dropdown = gr.Dropdown([r',', r'\t', r';', r' '], value=r',',
label='Delimiter',
allow_custom_value=True,
interactive=True)
utility_read_custom_delimiter = gr.Button('Read Tabular File')
utility_tabular_df = gr.DataFrame(label='Tabular Dataframe', interactive=False,
show_row_numbers=True, show_search='filter')
utility_read_custom_delimiter.click(lambda x, y: pd.read_csv(x, sep=y),
inputs=[utility_tabular_file,
utility_custom_delimiter_dropdown],
outputs=utility_tabular_df)
with gr.Accordion('Display CCD Ligand', open=False):
ccd_3d_ligand = gr.Text(label='CCD ID', interactive=True, info='Press Enter to submit')
with gr.Row():
ccd_3d_viewer = gr.HTML(get_ligand_molstar_html(''))
ccd_3d_info = gr.DataFrame(pd.DataFrame, headers=['Type', 'Program', 'Descriptor'])
ccd_3d_ligand.submit(draw_ccd_mol_3d, inputs=ccd_3d_ligand,
outputs=[ccd_3d_viewer, ccd_3d_info])
with gr.Accordion('Display SMILES Ligand', open=False):
smiles_3d_ligand = gr.Text(label='SMILES', interactive=True, info='Press Enter to submit')
with gr.Row():
smiles_3d_viewer = gr.HTML(get_ligand_molstar_html(''))
smiles_3d_info = gr.DataFrame(pd.DataFrame, headers=['Property', 'Value'],
column_widths=['80%', '20%'])
smiles_3d_ligand.submit(draw_smiles_3d, inputs=smiles_3d_ligand,
outputs=[smiles_3d_viewer, smiles_3d_info])
#####βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ#####
def change_sequence_label(curr_entity: str, sequence: str, cyclic_ckbox: bool):
cyclic_ckbox_bool = False if curr_entity in ['CCD', 'Ligand'] else True
return (gr.update(label=entity_label_map[curr_entity]),
validate_sequence(curr_entity, sequence),
gr.update(interactive=cyclic_ckbox_bool,
value=False if not cyclic_ckbox_bool else cyclic_ckbox),)
def validate_sequence(entity_type: str, sequence: str):
sequence = sequence.strip()
if not sequence:
return [('Input required!', "X")]
if entity_type in ["Protein", "DNA", "RNA"]:
sequence = sequence.upper()
labeled_sequence = []
prev_valid, prev_invalid = False, False
allowed_chars = allow_char_dict[entity_type]
for char in sequence:
if char not in allowed_chars:
if not prev_invalid:
labeled_sequence.append([char, "X"])
prev_valid = False
prev_invalid = True
else:
labeled_sequence[-1][0] += char
else:
if not prev_valid:
labeled_sequence.append([char, "β"])
prev_valid = True
prev_invalid = False
else:
labeled_sequence[-1][0] += char
if len(labeled_sequence) == 1 and prev_valid:
labeled_sequence = [('Valid', "β")]
elif entity_type == "Ligand":
mol = Chem.MolFromSmiles(sequence)
if mol is None:
labeled_sequence = [(sequence, "X")]
else:
# labeled_sequence = [(sequence, "β")]
labeled_sequence = [('Valid', "β")]
elif entity_type == 'CCD':
sequence = sequence.upper().strip()
if not re.fullmatch(r'(?:[A-Z0-9]{3}|[A-Z0-9]{5})|[A-Z]{2}', sequence):
labeled_sequence = [(sequence, "X")]
else:
# labeled_sequence = [(sequence, "β")]
labeled_sequence = [('Valid', "β")]
return labeled_sequence
def update_chain_seq_dict(entity_type: str, chain: str, seq: str, old_dict: dict):
if not all((chain, seq)):
return old_dict, gr.update(), gr.update()
old_dict.update({chain: {'type' : entity_type,
'sequence': seq,}})
return old_dict, gr.update(choices=list(old_dict)), gr.update(choices=list(old_dict))
if __name__ == '__main__':
import argparse
parser = argparse.ArgumentParser(description="Launch Boltz Gradio interface")
parser.add_argument("--share", action="store_true", help="Enable Gradio sharing (share=True)")
args = parser.parse_args()
threading.Thread(target=concurrent_download_model_weight, daemon=True).start()
Interface.launch(server_name="0.0.0.0", server_port=7860, share=args.share) |