Spaces:
Running
on
CPU Upgrade
Running
on
CPU Upgrade
Update app.py
Browse files
app.py
CHANGED
@@ -20,16 +20,22 @@ from Bio.Align import PairwiseAligner
|
|
20 |
import gradio as gr
|
21 |
import hydra
|
22 |
import pandas as pd
|
23 |
-
import plotly.express as px
|
24 |
import requests
|
25 |
-
from
|
26 |
from rdkit.Chem.rdMolDescriptors import CalcNumRotatableBonds, CalcNumHeavyAtoms, CalcNumAtoms, CalcTPSA
|
27 |
from requests.adapters import HTTPAdapter, Retry
|
28 |
from rdkit import Chem
|
29 |
-
from rdkit.Chem import RDConfig, Descriptors, Draw, Lipinski, Crippen, PandasTools
|
30 |
from rdkit.Chem.Scaffolds import MurckoScaffold
|
31 |
import seaborn as sns
|
32 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
33 |
import swifter
|
34 |
from tqdm.auto import tqdm
|
35 |
|
@@ -47,11 +53,11 @@ pd.set_option('display.float_format', '{:.3f}'.format)
|
|
47 |
PandasTools.molRepresentation = 'svg'
|
48 |
PandasTools.drawOptions = Draw.rdMolDraw2D.MolDrawOptions()
|
49 |
PandasTools.drawOptions.clearBackground = False
|
50 |
-
PandasTools.drawOptions.bondLineWidth = 1
|
51 |
PandasTools.drawOptions.explicitMethyl = True
|
52 |
PandasTools.drawOptions.singleColourWedgeBonds = True
|
53 |
PandasTools.drawOptions.useCDKAtomPalette()
|
54 |
-
PandasTools.molSize = (128,
|
55 |
|
56 |
SESSION = requests.Session()
|
57 |
ADAPTER = HTTPAdapter(max_retries=Retry(total=5, backoff_factor=0.1, status_forcelist=[500, 502, 503, 504]))
|
@@ -329,13 +335,13 @@ def rule_of_three(mol):
|
|
329 |
SCORE_MAP = {
|
330 |
'SAscore': sa_score,
|
331 |
'LogP': logp,
|
332 |
-
'Molecular
|
333 |
-
'Number of
|
334 |
-
'Molar
|
335 |
-
'H-
|
336 |
-
'H-Bond
|
337 |
-
'Rotatable
|
338 |
-
'Topological
|
339 |
}
|
340 |
|
341 |
FILTER_MAP = {
|
@@ -393,7 +399,6 @@ COLUMN_ALIASES = {
|
|
393 |
'ID2': 'Target ID',
|
394 |
'Y': 'Actual CPI/CPA',
|
395 |
'Y^': 'Predicted CPI/CPA',
|
396 |
-
'N': 'Original Index'
|
397 |
}
|
398 |
|
399 |
|
@@ -401,7 +406,7 @@ def validate_columns(df, mandatory_cols):
|
|
401 |
missing_cols = [col for col in mandatory_cols if col not in df.columns]
|
402 |
if missing_cols:
|
403 |
error_message = (f"The following mandatory columns are missing "
|
404 |
-
f"in the uploaded dataset: {str(
|
405 |
raise ValueError(error_message)
|
406 |
else:
|
407 |
return
|
@@ -540,17 +545,26 @@ def submit_predict(predict_filepath, task, preset, target_family, flag, state, p
|
|
540 |
|
541 |
def update_df(file, progress=gr.Progress(track_tqdm=True)):
|
542 |
# global DF_FOR_REPORT
|
543 |
-
if Path(file).is_file():
|
544 |
df = pd.read_csv(file)
|
|
|
|
|
|
|
|
|
|
|
545 |
# if df['X1'].nunique() > 1:
|
546 |
-
|
547 |
-
|
548 |
-
|
549 |
-
|
550 |
-
|
551 |
-
|
552 |
-
|
553 |
-
|
|
|
|
|
|
|
|
|
554 |
# DF_FOR_REPORT = df.copy()
|
555 |
|
556 |
# pie_chart = None
|
@@ -574,44 +588,81 @@ def update_df(file, progress=gr.Progress(track_tqdm=True)):
|
|
574 |
return {analyze_btn: gr.Button(interactive=False)}
|
575 |
|
576 |
|
577 |
-
def create_html_report(df, file=None, progress=gr.Progress(track_tqdm=True)):
|
578 |
df_html = df.copy(deep=True)
|
|
|
579 |
|
580 |
-
cols_left =
|
581 |
-
|
582 |
-
|
583 |
-
cols_right = [col for col in cols_right if col in df_html.columns]
|
584 |
df_html = df_html[cols_left + (df_html.columns.drop(cols_left + cols_right).tolist()) + cols_right]
|
585 |
|
586 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
587 |
df_html = df_html.sort_values(
|
588 |
[col for col in ['Y', 'Y^'] if col in df_html.columns], ascending=ascending
|
589 |
)
|
590 |
|
591 |
-
# # Remove repeated info for one-against-N tasks to save visual and physical space
|
592 |
-
# if df_html['X1'].nunique() <= 1:
|
593 |
-
# columns_to_clean = ['X1', 'ID1', 'Scaffold', 'Compound'] + list(FILTER_MAP.keys()) + list(SCORE_MAP.keys())
|
594 |
-
# for column in columns_to_clean:
|
595 |
-
# if column in df_html.columns:
|
596 |
-
# df_html.loc[1:, column] = pd.NA
|
597 |
-
#
|
598 |
-
# if df_html['X2'].nunique() <= 1:
|
599 |
-
# columns_to_clean = ['X2', 'ID2']
|
600 |
-
# for column in columns_to_clean:
|
601 |
-
# if column in df_html.columns:
|
602 |
-
# df_html.loc[1:, column] = pd.NA
|
603 |
-
|
604 |
if not file:
|
605 |
df_html = df_html.iloc[:31]
|
606 |
|
607 |
-
#
|
608 |
-
|
609 |
-
|
610 |
-
|
611 |
-
|
612 |
-
|
613 |
-
df_html
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
614 |
df_html.index.name = 'Index'
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
615 |
|
616 |
if not file:
|
617 |
if 'Compound ID' in df_html.columns:
|
@@ -620,100 +671,253 @@ def create_html_report(df, file=None, progress=gr.Progress(track_tqdm=True)):
|
|
620 |
df_html.drop(['Target FASTA'], axis=1, inplace=True)
|
621 |
if 'Target FASTA' in df_html.columns:
|
622 |
df_html['Target FASTA'] = df_html['Target FASTA'].swifter.progress_bar(
|
623 |
-
'Processing FASTA...').apply(
|
624 |
-
|
625 |
-
|
|
|
626 |
styled_df = df_html.style.format(precision=3)
|
627 |
-
colors = sns.color_palette('husl', len(df_html.columns))
|
628 |
-
for i, col in enumerate(df_html.columns):
|
629 |
-
if pd.api.types.is_numeric_dtype(df_html[col]):
|
630 |
-
styled_df = styled_df.background_gradient(subset=col, cmap=sns.light_palette(colors[i], as_cmap=True))
|
631 |
-
html = styled_df.to_html()
|
632 |
-
return f'Report preview<div style="overflow:auto; height: 300px; font-family: Courier !important;">{html}</div>'
|
633 |
-
else:
|
634 |
-
import panel as pn
|
635 |
-
from bokeh.resources import INLINE
|
636 |
-
from bokeh.models import NumberFormatter, BooleanFormatter
|
637 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
638 |
bool_formatters = {col: BooleanFormatter() for col in df_html.select_dtypes(bool).columns}
|
639 |
-
num_formatters = {col: NumberFormatter(format='0.000') for col in df_html.select_dtypes('
|
640 |
other_formatters = {
|
641 |
-
'Predicted
|
642 |
-
'Actual
|
643 |
-
'Compound': HTMLTemplateFormatter(),
|
644 |
-
'Scaffold': HTMLTemplateFormatter(),
|
645 |
'Target FASTA': {'type': 'textarea', 'width': 60},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
646 |
}
|
647 |
formatters = {**bool_formatters, **num_formatters, **other_formatters}
|
648 |
|
649 |
# html = df.to_html(file)
|
650 |
# return html
|
651 |
-
pn.widgets.Tabulator(df_html, formatters=formatters).save(file, resources=INLINE)
|
652 |
-
|
653 |
-
|
654 |
-
# def create_pie_chart(df, category, value, top_k):
|
655 |
-
# df.rename(COLUMN_ALIASES, inplace=True)
|
656 |
-
# # Select the top_k records based on the value_col
|
657 |
-
# top_k_df = df.nlargest(top_k, value)
|
658 |
-
#
|
659 |
-
# # Count the frequency of each unique value in the category_col column
|
660 |
-
# category_counts = top_k_df[category].value_counts()
|
661 |
-
#
|
662 |
-
# # Convert the counts to a DataFrame
|
663 |
-
# data = pd.DataFrame({category: category_counts.index, 'value': category_counts.values})
|
664 |
-
#
|
665 |
-
# # Calculate the angle for each category
|
666 |
-
# data['angle'] = data['value']/data['value'].sum() * 2*pi
|
667 |
-
#
|
668 |
-
# # Assign colors
|
669 |
-
# data['color'] = Spectral11[0:len(category_counts)]
|
670 |
-
#
|
671 |
-
# # Create the plot
|
672 |
-
# p = figure(height=350, title="Pie Chart", toolbar_location=None,
|
673 |
-
# tools="hover", tooltips="@{}: @value".format(category), x_range=(-0.5, 1.0))
|
674 |
-
#
|
675 |
-
# p.wedge(x=0, y=1, radius=0.4,
|
676 |
-
# start_angle=cumsum('angle', include_zero=True), end_angle=cumsum('angle'),
|
677 |
-
# line_color="white", fill_color='color', legend_field=category, source=data)
|
678 |
-
#
|
679 |
-
# p.axis.axis_label = None
|
680 |
-
# p.axis.visible = False
|
681 |
-
# p.grid.grid_line_color = None
|
682 |
-
#
|
683 |
-
# return p
|
684 |
|
685 |
-
|
686 |
-
|
687 |
-
|
688 |
-
|
689 |
-
# Select the top_k records based on the value_col
|
690 |
-
top_k_df = df.nlargest(top_k, value)
|
691 |
|
692 |
-
|
693 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
694 |
|
695 |
-
|
696 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
697 |
|
698 |
-
|
699 |
-
|
700 |
-
fig.update_traces(textposition='inside', textinfo='percent+label')
|
701 |
|
702 |
-
return fig
|
703 |
|
|
|
|
|
|
|
|
|
|
|
|
|
704 |
|
705 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
706 |
df_report = df.copy()
|
707 |
try:
|
708 |
for filter_name in filter_list:
|
709 |
df_report[filter_name] = df_report['Compound'].swifter.progress_bar(
|
710 |
desc=f"Calculating {filter_name}").apply(
|
711 |
-
lambda x: FILTER_MAP[filter_name](x) if not pd.isna(x) else x
|
712 |
|
713 |
for score_name in score_list:
|
714 |
df_report[score_name] = df_report['Compound'].swifter.progress_bar(
|
715 |
desc=f"Calculating {score_name}").apply(
|
716 |
-
lambda x: SCORE_MAP[score_name](x) if not pd.isna(x) else x
|
717 |
|
718 |
# pie_chart = None
|
719 |
# value = None
|
@@ -728,11 +932,13 @@ def submit_report(df, score_list, filter_list, progress=gr.Progress(track_tqdm=T
|
|
728 |
# elif df['X2'].nunique() > 1 >= df['X1'].nunique():
|
729 |
# pie_chart = create_pie_chart(df, category='Target family', value=value, top_k=100)
|
730 |
|
731 |
-
return create_html_report(df_report), df_report
|
|
|
732 |
|
733 |
except Exception as e:
|
734 |
gr.Warning(f'Failed to report results due to error: {str(e)}')
|
735 |
-
return None, None
|
|
|
736 |
|
737 |
# def check_job_status(job_id):
|
738 |
# job_lock = DATA_PATH / f"{job_id}.lock"
|
@@ -844,6 +1050,7 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
844 |
screen_flag = gr.State(value=False)
|
845 |
identify_flag = gr.State(value=False)
|
846 |
infer_flag = gr.State(value=False)
|
|
|
847 |
|
848 |
with gr.Tabs() as tabs:
|
849 |
with gr.TabItem(label='Drug Hit Screening', id=0):
|
@@ -916,7 +1123,8 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
916 |
|
917 |
with gr.Row():
|
918 |
with gr.Column():
|
919 |
-
target_family_detect_btn = gr.Button(value='OR Let Us Auto-Detect for You',
|
|
|
920 |
|
921 |
with gr.Row():
|
922 |
with gr.Column():
|
@@ -944,9 +1152,10 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
944 |
"while affinity prediction directly estimates their binding strength measured using "
|
945 |
"IC50."
|
946 |
)
|
947 |
-
drug_screen_task = gr.Dropdown(
|
948 |
-
|
949 |
-
|
|
|
950 |
|
951 |
with gr.Row():
|
952 |
with gr.Column():
|
@@ -954,10 +1163,11 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
954 |
"Select your preferred model, or click Recommend for the best-performing model based "
|
955 |
"on the selected task, family, and whether the target was trained. "
|
956 |
"Please refer to documentation for detailed benchamrk results."
|
957 |
-
|
958 |
drug_screen_preset = gr.Dropdown(list(PRESET_MAP.keys()),
|
959 |
label='Step 5. Select a Preset Model')
|
960 |
-
screen_preset_recommend_btn = gr.Button(
|
|
|
961 |
with gr.Row():
|
962 |
with gr.Column():
|
963 |
drug_screen_email = gr.Textbox(
|
@@ -1048,9 +1258,10 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1048 |
"while affinity prediction directly estimates their binding strength measured using "
|
1049 |
"IC50."
|
1050 |
)
|
1051 |
-
target_identify_task = gr.Dropdown(
|
1052 |
-
|
1053 |
-
|
|
|
1054 |
|
1055 |
with gr.Row():
|
1056 |
with gr.Column():
|
@@ -1058,7 +1269,7 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1058 |
"Select your preferred model, or click Recommend for the best-performing model based "
|
1059 |
"on the selected task, family, and whether the compound was trained. "
|
1060 |
"Please refer to documentation for detailed benchamrk results."
|
1061 |
-
|
1062 |
target_identify_preset = gr.Dropdown(list(PRESET_MAP.keys()),
|
1063 |
label='Step 5. Select a Preset Model')
|
1064 |
identify_preset_recommend_btn = gr.Button(value='OR Let Us Recommend for You',
|
@@ -1073,7 +1284,8 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1073 |
|
1074 |
with gr.Row(visible=True):
|
1075 |
# target_identify_clr_btn = gr.ClearButton(size='lg')
|
1076 |
-
target_identify_btn = gr.Button(value='SUBMIT THE IDENTIFICATION JOB', variant='primary',
|
|
|
1077 |
|
1078 |
identify_data_for_predict = gr.File(visible=False, file_count="single", type='filepath')
|
1079 |
identify_waiting = gr.Markdown(f"Your job is running... It might take a few minutes."
|
@@ -1152,9 +1364,10 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1152 |
"while affinity prediction directly estimates their binding strength "
|
1153 |
"measured using IC50."
|
1154 |
)
|
1155 |
-
pair_infer_task = gr.Dropdown(
|
1156 |
-
|
1157 |
-
|
|
|
1158 |
|
1159 |
with gr.Row():
|
1160 |
with gr.Column():
|
@@ -1189,17 +1402,20 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1189 |
To compute chemical properties for the predictions of drug hit screening,
|
1190 |
target protein identification, and interaction pair inference.
|
1191 |
|
1192 |
-
You may also upload your own dataset using a CSV file containing
|
|
|
1193 |
|
1194 |
The page shows only a preview report displaying at most 30 records
|
1195 |
(with top predicted CPI/CPA if reporting results from a prediction job).
|
1196 |
|
1197 |
-
|
1198 |
-
|
1199 |
-
|
1200 |
''')
|
1201 |
with gr.Row():
|
1202 |
-
|
|
|
|
|
|
|
1203 |
raw_df = gr.State(value=pd.DataFrame())
|
1204 |
report_df = gr.State(value=pd.DataFrame())
|
1205 |
scores = gr.CheckboxGroup(list(SCORE_MAP.keys()), label='Scores')
|
@@ -1207,7 +1423,8 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1207 |
|
1208 |
with gr.Row():
|
1209 |
# clear_btn = gr.ClearButton(size='lg')
|
1210 |
-
analyze_btn = gr.Button('Preview Top 30 Records', variant='primary', size='lg',
|
|
|
1211 |
|
1212 |
with gr.Row():
|
1213 |
with gr.Column(scale=3):
|
@@ -1217,11 +1434,11 @@ with gr.Blocks(theme=theme, title='DeepSEQreen', css=CSS) as demo:
|
|
1217 |
with gr.Row():
|
1218 |
with gr.Column():
|
1219 |
csv_generate = gr.Button(value='Generate CSV Report',
|
1220 |
-
interactive=
|
1221 |
csv_download_file = gr.File(label='Download CSV Report', visible=False)
|
1222 |
with gr.Column():
|
1223 |
html_generate = gr.Button(value='Generate HTML Report',
|
1224 |
-
interactive=
|
1225 |
html_download_file = gr.File(label='Download HTML Report', visible=False)
|
1226 |
|
1227 |
|
@@ -1336,6 +1553,8 @@ QALAHAYFAQYHDPDDEPVADPYDQSFESRDLLIDEWKSLTYDEVISFVPPPLDQEEMES
|
|
1336 |
|
1337 |
example_fasta.click(fn=example_fill, inputs=target_input_type, outputs=[
|
1338 |
target_id, target_gene, target_organism, target_fasta], show_progress=False)
|
|
|
|
|
1339 |
# example_uniprot.click(fn=example_fill, inputs=target_input_type, outputs=target_fasta, show_progress=False)
|
1340 |
# example_gene.click(fn=example_fill, inputs=target_input_type, outputs=target_fasta, show_progress=False)
|
1341 |
|
@@ -1663,47 +1882,82 @@ QALAHAYFAQYHDPDDEPVADPYDQSFESRDLLIDEWKSLTYDEVISFVPPPLDQEEMES
|
|
1663 |
)
|
1664 |
|
1665 |
# TODO background job from these 3 pipelines to update file_for_report
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1666 |
|
1667 |
file_for_report.change(fn=update_df, inputs=file_for_report, outputs=[
|
1668 |
-
html_report,
|
1669 |
-
|
1670 |
-
|
1671 |
-
|
1672 |
-
|
1673 |
-
|
1674 |
-
|
1675 |
-
|
1676 |
-
|
1677 |
-
|
1678 |
-
|
1679 |
-
|
1680 |
-
|
1681 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1682 |
try:
|
1683 |
now = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
1684 |
filename = f"reports/{Path(file_report.name).stem}_DeepSEQreen_report_{now}.csv"
|
1685 |
df.drop(labels=['Compound', 'Scaffold'], axis=1).to_csv(filename, index=False)
|
1686 |
|
1687 |
-
return gr.File(filename
|
1688 |
except Exception as e:
|
1689 |
gr.Warning(f"Failed to generate CSV due to error: {str(e)}")
|
1690 |
-
return None
|
|
|
1691 |
|
1692 |
-
def create_html_report_file(df, file_report):
|
1693 |
try:
|
1694 |
now = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
1695 |
filename = f"reports/{Path(file_report.name).stem}_DeepSEQreen_report_{now}.html"
|
1696 |
create_html_report(df, filename)
|
1697 |
-
return gr.File(filename, visible=True)
|
1698 |
except Exception as e:
|
1699 |
gr.Warning(f"Failed to generate HTML due to error: {str(e)}")
|
1700 |
-
return None
|
|
|
1701 |
|
1702 |
html_report.change(lambda: [gr.Button(visible=True)] * 2, outputs=[csv_generate, html_generate])
|
1703 |
-
csv_generate.click(
|
1704 |
-
|
1705 |
-
|
1706 |
-
|
|
|
|
|
|
|
|
|
1707 |
|
1708 |
# screen_waiting.change(fn=check_job_status, inputs=run_state, outputs=[pair_waiting, tabs, file_for_report],
|
1709 |
# every=5)
|
|
|
20 |
import gradio as gr
|
21 |
import hydra
|
22 |
import pandas as pd
|
|
|
23 |
import requests
|
24 |
+
from rdkit.Chem.PandasTools import _MolPlusFingerprint
|
25 |
from rdkit.Chem.rdMolDescriptors import CalcNumRotatableBonds, CalcNumHeavyAtoms, CalcNumAtoms, CalcTPSA
|
26 |
from requests.adapters import HTTPAdapter, Retry
|
27 |
from rdkit import Chem
|
28 |
+
from rdkit.Chem import RDConfig, Descriptors, Draw, Lipinski, Crippen, PandasTools
|
29 |
from rdkit.Chem.Scaffolds import MurckoScaffold
|
30 |
import seaborn as sns
|
31 |
|
32 |
+
from bokeh.models import Legend, NumberFormatter, BooleanFormatter, HTMLTemplateFormatter, LegendItem
|
33 |
+
from bokeh.palettes import Category20c_20
|
34 |
+
from bokeh.plotting import figure
|
35 |
+
from bokeh.transform import cumsum
|
36 |
+
from bokeh.resources import INLINE
|
37 |
+
import panel as pn
|
38 |
+
|
39 |
import swifter
|
40 |
from tqdm.auto import tqdm
|
41 |
|
|
|
53 |
PandasTools.molRepresentation = 'svg'
|
54 |
PandasTools.drawOptions = Draw.rdMolDraw2D.MolDrawOptions()
|
55 |
PandasTools.drawOptions.clearBackground = False
|
56 |
+
PandasTools.drawOptions.bondLineWidth = 1
|
57 |
PandasTools.drawOptions.explicitMethyl = True
|
58 |
PandasTools.drawOptions.singleColourWedgeBonds = True
|
59 |
PandasTools.drawOptions.useCDKAtomPalette()
|
60 |
+
PandasTools.molSize = (128, 80)
|
61 |
|
62 |
SESSION = requests.Session()
|
63 |
ADAPTER = HTTPAdapter(max_retries=Retry(total=5, backoff_factor=0.1, status_forcelist=[500, 502, 503, 504]))
|
|
|
335 |
SCORE_MAP = {
|
336 |
'SAscore': sa_score,
|
337 |
'LogP': logp,
|
338 |
+
'Molecular Weight': mw,
|
339 |
+
'Number of Heavy Atoms': heavy_atom,
|
340 |
+
'Molar Refractivity': mr,
|
341 |
+
'H-Bond Donor Count': hbd,
|
342 |
+
'H-Bond Acceptor Count': hba,
|
343 |
+
'Rotatable Bond Count': rotatable_bond,
|
344 |
+
'Topological Polar Surface Area': tpsa,
|
345 |
}
|
346 |
|
347 |
FILTER_MAP = {
|
|
|
399 |
'ID2': 'Target ID',
|
400 |
'Y': 'Actual CPI/CPA',
|
401 |
'Y^': 'Predicted CPI/CPA',
|
|
|
402 |
}
|
403 |
|
404 |
|
|
|
406 |
missing_cols = [col for col in mandatory_cols if col not in df.columns]
|
407 |
if missing_cols:
|
408 |
error_message = (f"The following mandatory columns are missing "
|
409 |
+
f"in the uploaded dataset: {str(mandatory_cols).strip('[]')}.")
|
410 |
raise ValueError(error_message)
|
411 |
else:
|
412 |
return
|
|
|
545 |
|
546 |
def update_df(file, progress=gr.Progress(track_tqdm=True)):
|
547 |
# global DF_FOR_REPORT
|
548 |
+
if file and Path(file).is_file():
|
549 |
df = pd.read_csv(file)
|
550 |
+
if 'N' in df.columns:
|
551 |
+
df.set_index('N', inplace=True)
|
552 |
+
if not any(col in ['X1', 'X2'] for col in df.columns):
|
553 |
+
gr.Warning("At least one of columns `X1` and `X2` must be in the uploaded dataset.")
|
554 |
+
return {analyze_btn: gr.Button(interactive=False)}
|
555 |
# if df['X1'].nunique() > 1:
|
556 |
+
if 'X1' in df.columns:
|
557 |
+
df['Scaffold SMILES'] = df['X1'].swifter.progress_bar(
|
558 |
+
desc=f"Calculating scaffold...").apply(MurckoScaffold.MurckoScaffoldSmilesFromSmiles)
|
559 |
+
df['Scaffold'] = df['Scaffold SMILES'].swifter.progress_bar(
|
560 |
+
desc='Generating scaffold graphs...').apply(
|
561 |
+
lambda smiles: _MolPlusFingerprint(Chem.MolFromSmiles(smiles)))
|
562 |
+
# Add a new column with RDKit molecule objects
|
563 |
+
if 'Compound' not in df.columns or df['Compound'].dtype != 'object':
|
564 |
+
df['Compound'] = df['X1'].swifter.progress_bar(
|
565 |
+
desc='Generating molecular graphs...').apply(
|
566 |
+
lambda smiles: _MolPlusFingerprint(Chem.MolFromSmiles(smiles)))
|
567 |
+
|
568 |
# DF_FOR_REPORT = df.copy()
|
569 |
|
570 |
# pie_chart = None
|
|
|
588 |
return {analyze_btn: gr.Button(interactive=False)}
|
589 |
|
590 |
|
591 |
+
def create_html_report(df, file=None, task=None, progress=gr.Progress(track_tqdm=True)):
|
592 |
df_html = df.copy(deep=True)
|
593 |
+
# email_hash = hashlib.sha256(email.encode()).hexdigest()
|
594 |
|
595 |
+
cols_left = list(pd.Index(
|
596 |
+
['ID1', 'Compound', 'Scaffold', 'Scaffold SMILES', 'ID2', 'Y', 'Y^']).intersection(df_html.columns))
|
597 |
+
cols_right = list(pd.Index(['X1', 'X2']).intersection(df_html.columns))
|
|
|
598 |
df_html = df_html[cols_left + (df_html.columns.drop(cols_left + cols_right).tolist()) + cols_right]
|
599 |
|
600 |
+
if isinstance(task, str):
|
601 |
+
task = TASK_MAP[task]
|
602 |
+
COLUMN_ALIASES.update({
|
603 |
+
'Y': 'Actual Interaction Probability' if task == 'DTI' else 'Actual Binding Affinity',
|
604 |
+
'Y^': 'Predicted Interaction Probability' if task == 'DTI' else 'Predicted Binding Affinity'
|
605 |
+
})
|
606 |
+
|
607 |
+
ascending = True if COLUMN_ALIASES['Y^'] == 'Predicted Binding Affinity' else False
|
608 |
df_html = df_html.sort_values(
|
609 |
[col for col in ['Y', 'Y^'] if col in df_html.columns], ascending=ascending
|
610 |
)
|
611 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
612 |
if not file:
|
613 |
df_html = df_html.iloc[:31]
|
614 |
|
615 |
+
# Remove repeated info for one-against-N tasks to save visual and physical space
|
616 |
+
job = 'Chemical Property'
|
617 |
+
unique_entity = 'Unique Entity'
|
618 |
+
unique_df = None
|
619 |
+
category = None
|
620 |
+
columns_unique = None
|
621 |
+
if 'X1' in df_html.columns and 'X2' in df_html.columns:
|
622 |
+
n_compound = df_html['X1'].nunique()
|
623 |
+
n_protein = df_html['X2'].nunique()
|
624 |
+
|
625 |
+
if n_compound == 1 and n_protein >= 2:
|
626 |
+
unique_entity = 'Compound of Interest'
|
627 |
+
if any(col in df_html.columns for col in ['Y^', 'Y']):
|
628 |
+
job = 'Target Protein Identification'
|
629 |
+
category = 'Target Family'
|
630 |
+
columns_unique = df_html.columns.isin(['X1', 'ID1', 'Scaffold', 'Compound', 'Scaffold SMILES']
|
631 |
+
+ list(FILTER_MAP.keys()) + list(SCORE_MAP.keys()))
|
632 |
+
|
633 |
+
elif n_compound >= 2 and n_protein == 1:
|
634 |
+
unique_entity = 'Target of Interest'
|
635 |
+
if any(col in df_html.columns for col in ['Y^', 'Y']):
|
636 |
+
job = 'Drug Hit Screening'
|
637 |
+
category = 'Scaffold SMILES'
|
638 |
+
columns_unique = df_html.columns.isin(['X2', 'ID2'])
|
639 |
+
|
640 |
+
elif 'Y^' in df_html.columns:
|
641 |
+
job = 'Interaction Pair Inference'
|
642 |
+
if 'Compound' in df_html.columns:
|
643 |
+
df_html['Compound'] = df_html['Compound'].swifter.progress_bar(
|
644 |
+
desc='Generating compound graph...').apply(
|
645 |
+
lambda x: PandasTools.PrintAsImageString(x) if not pd.isna(x) else x)
|
646 |
+
if 'Scaffold' in df_html.columns:
|
647 |
+
df_html['Scaffold'] = df_html['Scaffold'].swifter.progress_bar(
|
648 |
+
desc='Generating scaffold graph...').apply(
|
649 |
+
lambda x: PandasTools.PrintAsImageString(x) if not pd.isna(x) else x)
|
650 |
+
|
651 |
+
df_html.rename(columns=COLUMN_ALIASES, inplace=True)
|
652 |
df_html.index.name = 'Index'
|
653 |
+
if 'Target FASTA' in df_html.columns:
|
654 |
+
df_html['Target FASTA'] = df_html['Target FASTA'].swifter.progress_bar(
|
655 |
+
desc='Processing FASTA...').apply(
|
656 |
+
lambda x: wrap_text(x) if not pd.isna(x) else x)
|
657 |
+
|
658 |
+
if columns_unique is not None:
|
659 |
+
unique_df = df_html.loc[:, columns_unique].iloc[[0]]
|
660 |
+
df_html = df_html.loc[:, ~columns_unique]
|
661 |
+
|
662 |
+
num_cols = df_html.select_dtypes('number').columns
|
663 |
+
num_col_colors = sns.color_palette('husl', len(num_cols))
|
664 |
+
bool_cols = df_html.select_dtypes(bool).columns
|
665 |
+
bool_col_colors = {True: 'lightgreen', False: 'lightpink'}
|
666 |
|
667 |
if not file:
|
668 |
if 'Compound ID' in df_html.columns:
|
|
|
671 |
df_html.drop(['Target FASTA'], axis=1, inplace=True)
|
672 |
if 'Target FASTA' in df_html.columns:
|
673 |
df_html['Target FASTA'] = df_html['Target FASTA'].swifter.progress_bar(
|
674 |
+
desc='Processing FASTA...').apply(
|
675 |
+
lambda x: wrap_text(x) if not pd.isna(x) else x)
|
676 |
+
if 'Scaffold SMILES' in df_html.columns:
|
677 |
+
df_html.drop(['Scaffold SMILES'], axis=1, inplace=True)
|
678 |
styled_df = df_html.style.format(precision=3)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
679 |
|
680 |
+
for i, col in enumerate(num_cols):
|
681 |
+
if col not in ['Predicted Binding Affinity', 'Actual Binding Affinity']:
|
682 |
+
styled_df = styled_df.background_gradient(
|
683 |
+
subset=[col], cmap=sns.light_palette(num_col_colors[i], as_cmap=True))
|
684 |
+
else:
|
685 |
+
styled_df = styled_df.background_gradient(
|
686 |
+
subset=[col], cmap=sns.light_palette(num_col_colors[i], as_cmap=True).reversed())
|
687 |
+
|
688 |
+
styled_df.applymap(lambda val: f'background-color: {bool_col_colors[val]}', subset=bool_cols)
|
689 |
+
|
690 |
+
table_html = styled_df.to_html()
|
691 |
+
unique_html = ''
|
692 |
+
if unique_df is not None:
|
693 |
+
unique_html = unique_df.replace('\n', '<br>', regex=True).to_html(escape=False, index=False)
|
694 |
+
unique_html = f'<div style="font-family: Courier !important;">{unique_html}</div>'
|
695 |
+
|
696 |
+
return (f'<div style="font-size: 16px; font-weight: bold;">{job} Report Preview (Top 30 Records)</div>'
|
697 |
+
f'{unique_html}'
|
698 |
+
f'<div style="overflow:auto; height: 300px; font-family: Courier !important;">{table_html}</div>')
|
699 |
+
|
700 |
+
else:
|
701 |
bool_formatters = {col: BooleanFormatter() for col in df_html.select_dtypes(bool).columns}
|
702 |
+
num_formatters = {col: NumberFormatter(format='0.000') for col in df_html.select_dtypes('floating').columns}
|
703 |
other_formatters = {
|
704 |
+
'Predicted Interaction Probability': {'type': 'progress', 'max': 1.0, 'legend': True},
|
705 |
+
'Actual Interaction Probability': {'type': 'progress', 'max': 1.0, 'legend': True},
|
706 |
+
'Compound': HTMLTemplateFormatter(template='<div class="image-zoom-viewer"><%= value %></div>'),
|
707 |
+
'Scaffold': HTMLTemplateFormatter(template='<div class="image-zoom-viewer"><%= value %></div>'),
|
708 |
'Target FASTA': {'type': 'textarea', 'width': 60},
|
709 |
+
'Target ID': HTMLTemplateFormatter(
|
710 |
+
template='<a href="<% '
|
711 |
+
'if (/^[OPQ][0-9][A-Z0-9]{3}[0-9]|[A-NR-Z][0-9]([A-Z][A-Z0-9]{2}[0-9]){1,2}$/.test(value)) '
|
712 |
+
'{ %>https://www.uniprot.org/uniprotkb/<%= value %><% } '
|
713 |
+
'else { %>https://www.uniprot.org/uniprotkb?query=<%= value %><% } '
|
714 |
+
'%>" target="_blank"><%= value %></a>'),
|
715 |
+
'Compound ID': HTMLTemplateFormatter(
|
716 |
+
template='<a href="https://pubchem.ncbi.nlm.nih.gov/compound/<%= value %>" '
|
717 |
+
'target="_blank"><%= value %></a>')
|
718 |
}
|
719 |
formatters = {**bool_formatters, **num_formatters, **other_formatters}
|
720 |
|
721 |
# html = df.to_html(file)
|
722 |
# return html
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
723 |
|
724 |
+
report_table = pn.widgets.Tabulator(
|
725 |
+
df_html, formatters=formatters,
|
726 |
+
frozen_columns=['Index', 'Target ID', 'Compound ID', 'Compound', 'Scaffold'],
|
727 |
+
disabled=True, sizing_mode='stretch_both')
|
|
|
|
|
728 |
|
729 |
+
for i, col in enumerate(num_cols):
|
730 |
+
if col not in ['Predicted Binding Affinity', 'Actual Binding Affinity']:
|
731 |
+
if col not in ['Predicted Interaction Probability', 'Actual Interaction Probability']:
|
732 |
+
report_table.style.background_gradient(
|
733 |
+
subset=df_html.columns == col, cmap=sns.light_palette(num_col_colors[i], as_cmap=True))
|
734 |
+
else:
|
735 |
+
continue
|
736 |
+
else:
|
737 |
+
report_table.style.background_gradient(
|
738 |
+
subset=df_html.columns == col, cmap=sns.light_palette(num_col_colors[i], as_cmap=True).reversed())
|
739 |
+
|
740 |
+
pie_charts = {}
|
741 |
+
for y in df_html.columns.intersection(['Predicted Interaction Probability', 'Actual Interaction Probability',
|
742 |
+
'Predicted Binding Affinity', 'Actual Binding Affinity']):
|
743 |
+
pie_charts[y] = []
|
744 |
+
for k in [10, 30, 100]:
|
745 |
+
if k < len(df_html):
|
746 |
+
pie_charts[y].append(create_pie_chart(df_html, category=category, value=y, top_k=k))
|
747 |
+
pie_charts[y].append(create_pie_chart(df_html, category=category, value=y, top_k=len(df_html)))
|
748 |
+
|
749 |
+
# Remove keys with empty values
|
750 |
+
pie_charts = {k: v for k, v in pie_charts.items() if any(v)}
|
751 |
+
|
752 |
+
pn_css = """
|
753 |
+
.tabulator {
|
754 |
+
font-family: Courier New !important;
|
755 |
+
font-weight: normal !important;
|
756 |
+
font-size: 12px !important;
|
757 |
+
overflow: visible !important;
|
758 |
+
}
|
759 |
+
|
760 |
+
.tabulator-cell {
|
761 |
+
overflow: visible !important;
|
762 |
+
}
|
763 |
+
|
764 |
+
.bk-panel-models-tabulator-DataTabulator {
|
765 |
+
overflow: visible !important;
|
766 |
+
}
|
767 |
+
|
768 |
+
.tabulator-cell.tabulator-frozen:hover {
|
769 |
+
z-index: 1000 !important;
|
770 |
+
}
|
771 |
+
|
772 |
+
.bk-panel-models-tabulator-DataTabulator:hover {
|
773 |
+
z-index: 999 !important;
|
774 |
+
}
|
775 |
+
|
776 |
+
.image-zoom-viewer {
|
777 |
+
display: inline-block;
|
778 |
+
position: relative;
|
779 |
+
overflow: visible; /* Ensures that the scaled SVG isn't clipped */
|
780 |
+
}
|
781 |
+
|
782 |
+
.image-zoom-viewer::after {
|
783 |
+
content: "";
|
784 |
+
position: absolute;
|
785 |
+
top: 0;
|
786 |
+
left: 0;
|
787 |
+
width: 100%;
|
788 |
+
height: 100%;
|
789 |
+
pointer-events: none;
|
790 |
+
}
|
791 |
+
|
792 |
+
.image-zoom-viewer:hover::after {
|
793 |
+
pointer-events: all;
|
794 |
+
}
|
795 |
+
|
796 |
+
/* When hovering over the container, scale its child (the SVG) */
|
797 |
+
.tabulator-cell:hover .image-zoom-viewer svg {
|
798 |
+
padding: 3px;
|
799 |
+
position: relative; /* Position the SVG relative to the viewport */
|
800 |
+
background-color: rgba(250, 250, 250, 0.854);
|
801 |
+
box-shadow: 0 0 10px rgba(0, 0, 0, 0.618);
|
802 |
+
border-radius: 3px;
|
803 |
+
transform: scale(4) translate(+38.2%, +38.2%); /* Scale up the SVG */
|
804 |
+
transition: transform 0.3s ease;
|
805 |
+
pointer-events: none; /* Prevents the SVG from blocking mouse interactions */
|
806 |
+
}
|
807 |
+
|
808 |
+
.image-zoom-viewer svg {
|
809 |
+
display: block; /* SVG is a block-level element for proper scaling */
|
810 |
+
z-index: 1000;
|
811 |
+
}
|
812 |
+
|
813 |
+
.image-zoom-viewer:hover {
|
814 |
+
z-index: 1000;
|
815 |
+
}
|
816 |
+
|
817 |
+
"""
|
818 |
+
|
819 |
+
pn.extension(raw_css=[pn_css])
|
820 |
+
|
821 |
+
template = pn.template.VanillaTemplate(
|
822 |
+
title=f'DeepSEQreen {job} Report',
|
823 |
+
sidebar=[],
|
824 |
+
favicon='deepseqreen.svg',
|
825 |
+
logo='deepseqreen.svg',
|
826 |
+
header_background='#F3F5F7',
|
827 |
+
header_color='#4372c4',
|
828 |
+
busy_indicator=None,
|
829 |
+
)
|
830 |
|
831 |
+
info_row = pn.Row()
|
832 |
+
if unique_df is not None:
|
833 |
+
unique_table = pn.widgets.Tabulator(unique_df, formatters=formatters, show_index=False, disabled=True)
|
834 |
+
info_row.append(pn.Column(f'### {unique_entity}', unique_table))
|
835 |
+
if pie_charts:
|
836 |
+
for score_name, figure_list in pie_charts.items():
|
837 |
+
info_row.append(
|
838 |
+
pn.Column(f'### {category} by Top {score_name}',
|
839 |
+
pn.Tabs(*figure_list, tabs_location='above'))
|
840 |
+
# pn.Card(pn.Row(v), title=f'{category} by Top {k}')
|
841 |
+
)
|
842 |
+
if info_row:
|
843 |
+
template.main.append(pn.Card(info_row,
|
844 |
+
sizing_mode='stretch_width', title='Summary Statistics', margin=10))
|
845 |
+
|
846 |
+
template.main.append(
|
847 |
+
pn.Card(report_table, title=f'{job} Results', # width=1200,
|
848 |
+
margin=10)
|
849 |
+
)
|
850 |
|
851 |
+
template.save(file, resources=INLINE)
|
852 |
+
return file
|
|
|
853 |
|
|
|
854 |
|
855 |
+
def create_pie_chart(df, category, value, top_k):
|
856 |
+
if category not in df or value not in df:
|
857 |
+
return
|
858 |
+
top_k_df = df.nlargest(top_k, value)
|
859 |
+
category_counts = top_k_df[category].value_counts()
|
860 |
+
data = pd.DataFrame({category: category_counts.index, 'value': category_counts.values})
|
861 |
|
862 |
+
data['proportion'] = data['value'] / data['value'].sum()
|
863 |
+
# Merge rows with proportion less than 0.1% into one row
|
864 |
+
mask = data['proportion'] <= 0.001
|
865 |
+
merged_row = data[mask].sum()
|
866 |
+
merged_row[category] = 'Other'
|
867 |
+
data = pd.concat([data[~mask], pd.DataFrame(merged_row).T])
|
868 |
+
data['angle'] = data['proportion'] * 2 * pi
|
869 |
+
data['color'] = (Category20c_20 * (len(data) // 20 + 1))[:len(data)]
|
870 |
+
|
871 |
+
tooltips = [
|
872 |
+
(f"{category}", f"@{{{category}}}"),
|
873 |
+
("Count", "@value"),
|
874 |
+
("Percentage", "@proportion{0.0%}")
|
875 |
+
]
|
876 |
+
|
877 |
+
if category == 'Scaffold SMILES':
|
878 |
+
data = data.merge(top_k_df[['Scaffold SMILES', 'Scaffold']].drop_duplicates(), how='left',
|
879 |
+
left_on='Scaffold SMILES', right_on='Scaffold SMILES')
|
880 |
+
tooltips.append(("Scaffold", "<div>@{Scaffold}{safe}</div>"))
|
881 |
+
p = figure(height=256, name=f"Top {top_k}" if top_k < len(df) else 'All',
|
882 |
+
toolbar_location=None, tools="hover", tooltips=tooltips, x_range=(-0.5, 0.5),
|
883 |
+
sizing_mode="scale_height")
|
884 |
+
p.axis.axis_label = None
|
885 |
+
p.axis.visible = False
|
886 |
+
p.grid.grid_line_color = None
|
887 |
+
p.outline_line_width = 0
|
888 |
+
p.min_border = 0
|
889 |
+
p.min_border_right = 0
|
890 |
+
p.margin = 0
|
891 |
+
|
892 |
+
p.add_layout(Legend(padding=0, margin=0), 'right')
|
893 |
+
p.wedge(x=0, y=1, radius=0.3,
|
894 |
+
start_angle=cumsum('angle', include_zero=True), end_angle=cumsum('angle'),
|
895 |
+
line_color="white", fill_color='color', legend_field=category, source=data)
|
896 |
+
|
897 |
+
p.legend.label_text_font_size = "8pt"
|
898 |
+
p.legend.margin = 0
|
899 |
+
p.legend.padding = 0
|
900 |
+
|
901 |
+
# Limit the number of legend items to 20 and add "..." if there are more than 30 items
|
902 |
+
if len(p.legend.items) > 20:
|
903 |
+
p.legend.items = p.legend.items[:21]
|
904 |
+
p.legend.items.append(LegendItem(label="..."))
|
905 |
+
|
906 |
+
return p
|
907 |
+
|
908 |
+
|
909 |
+
def submit_report(df, score_list, filter_list, task, progress=gr.Progress(track_tqdm=True)):
|
910 |
df_report = df.copy()
|
911 |
try:
|
912 |
for filter_name in filter_list:
|
913 |
df_report[filter_name] = df_report['Compound'].swifter.progress_bar(
|
914 |
desc=f"Calculating {filter_name}").apply(
|
915 |
+
lambda x: FILTER_MAP[filter_name](x) if not pd.isna(x) else x)
|
916 |
|
917 |
for score_name in score_list:
|
918 |
df_report[score_name] = df_report['Compound'].swifter.progress_bar(
|
919 |
desc=f"Calculating {score_name}").apply(
|
920 |
+
lambda x: SCORE_MAP[score_name](x) if not pd.isna(x) else x)
|
921 |
|
922 |
# pie_chart = None
|
923 |
# value = None
|
|
|
932 |
# elif df['X2'].nunique() > 1 >= df['X1'].nunique():
|
933 |
# pie_chart = create_pie_chart(df, category='Target family', value=value, top_k=100)
|
934 |
|
935 |
+
return (create_html_report(df_report, file=None, task=task), df_report,
|
936 |
+
gr.File(visible=False), gr.File(visible=False))
|
937 |
|
938 |
except Exception as e:
|
939 |
gr.Warning(f'Failed to report results due to error: {str(e)}')
|
940 |
+
return None, None, None, None
|
941 |
+
|
942 |
|
943 |
# def check_job_status(job_id):
|
944 |
# job_lock = DATA_PATH / f"{job_id}.lock"
|
|
|
1050 |
screen_flag = gr.State(value=False)
|
1051 |
identify_flag = gr.State(value=False)
|
1052 |
infer_flag = gr.State(value=False)
|
1053 |
+
report_upload_flag = gr.State(value=False)
|
1054 |
|
1055 |
with gr.Tabs() as tabs:
|
1056 |
with gr.TabItem(label='Drug Hit Screening', id=0):
|
|
|
1123 |
|
1124 |
with gr.Row():
|
1125 |
with gr.Column():
|
1126 |
+
target_family_detect_btn = gr.Button(value='OR Let Us Auto-Detect for You',
|
1127 |
+
variant='primary')
|
1128 |
|
1129 |
with gr.Row():
|
1130 |
with gr.Column():
|
|
|
1152 |
"while affinity prediction directly estimates their binding strength measured using "
|
1153 |
"IC50."
|
1154 |
)
|
1155 |
+
drug_screen_task = gr.Dropdown(
|
1156 |
+
list(TASK_MAP.keys()),
|
1157 |
+
label='Step 4. Select the Prediction Task You Want to Conduct',
|
1158 |
+
value='Compound-protein interaction')
|
1159 |
|
1160 |
with gr.Row():
|
1161 |
with gr.Column():
|
|
|
1163 |
"Select your preferred model, or click Recommend for the best-performing model based "
|
1164 |
"on the selected task, family, and whether the target was trained. "
|
1165 |
"Please refer to documentation for detailed benchamrk results."
|
1166 |
+
)
|
1167 |
drug_screen_preset = gr.Dropdown(list(PRESET_MAP.keys()),
|
1168 |
label='Step 5. Select a Preset Model')
|
1169 |
+
screen_preset_recommend_btn = gr.Button(
|
1170 |
+
value='OR Let Us Recommend for You', variant='primary')
|
1171 |
with gr.Row():
|
1172 |
with gr.Column():
|
1173 |
drug_screen_email = gr.Textbox(
|
|
|
1258 |
"while affinity prediction directly estimates their binding strength measured using "
|
1259 |
"IC50."
|
1260 |
)
|
1261 |
+
target_identify_task = gr.Dropdown(
|
1262 |
+
list(TASK_MAP.keys()),
|
1263 |
+
label='Step 4. Select the Prediction Task You Want to Conduct',
|
1264 |
+
value='Compound-protein interaction')
|
1265 |
|
1266 |
with gr.Row():
|
1267 |
with gr.Column():
|
|
|
1269 |
"Select your preferred model, or click Recommend for the best-performing model based "
|
1270 |
"on the selected task, family, and whether the compound was trained. "
|
1271 |
"Please refer to documentation for detailed benchamrk results."
|
1272 |
+
)
|
1273 |
target_identify_preset = gr.Dropdown(list(PRESET_MAP.keys()),
|
1274 |
label='Step 5. Select a Preset Model')
|
1275 |
identify_preset_recommend_btn = gr.Button(value='OR Let Us Recommend for You',
|
|
|
1284 |
|
1285 |
with gr.Row(visible=True):
|
1286 |
# target_identify_clr_btn = gr.ClearButton(size='lg')
|
1287 |
+
target_identify_btn = gr.Button(value='SUBMIT THE IDENTIFICATION JOB', variant='primary',
|
1288 |
+
size='lg')
|
1289 |
|
1290 |
identify_data_for_predict = gr.File(visible=False, file_count="single", type='filepath')
|
1291 |
identify_waiting = gr.Markdown(f"Your job is running... It might take a few minutes."
|
|
|
1364 |
"while affinity prediction directly estimates their binding strength "
|
1365 |
"measured using IC50."
|
1366 |
)
|
1367 |
+
pair_infer_task = gr.Dropdown(
|
1368 |
+
list(TASK_MAP.keys()),
|
1369 |
+
label='Step 3. Select the Prediction Task You Want to Conduct',
|
1370 |
+
value='Compound-protein interaction')
|
1371 |
|
1372 |
with gr.Row():
|
1373 |
with gr.Column():
|
|
|
1402 |
To compute chemical properties for the predictions of drug hit screening,
|
1403 |
target protein identification, and interaction pair inference.
|
1404 |
|
1405 |
+
You may also upload your own dataset using a CSV file containing
|
1406 |
+
one required column `X1` for compound SMILES.
|
1407 |
|
1408 |
The page shows only a preview report displaying at most 30 records
|
1409 |
(with top predicted CPI/CPA if reporting results from a prediction job).
|
1410 |
|
1411 |
+
Please first `**Preview**` the report, then `**Generate**` and download a CSV report
|
1412 |
+
or an interactive HTML report below if you wish to access the full report.
|
|
|
1413 |
''')
|
1414 |
with gr.Row():
|
1415 |
+
with gr.Column():
|
1416 |
+
file_for_report = gr.File(interactive=True, type='filepath')
|
1417 |
+
report_task = gr.Dropdown(list(TASK_MAP.keys()), visible=False, value=None,
|
1418 |
+
label='Specify the Task for the Labels in the Upload Dataset')
|
1419 |
raw_df = gr.State(value=pd.DataFrame())
|
1420 |
report_df = gr.State(value=pd.DataFrame())
|
1421 |
scores = gr.CheckboxGroup(list(SCORE_MAP.keys()), label='Scores')
|
|
|
1423 |
|
1424 |
with gr.Row():
|
1425 |
# clear_btn = gr.ClearButton(size='lg')
|
1426 |
+
analyze_btn = gr.Button('Preview Top 30 Records', variant='primary', size='lg',
|
1427 |
+
interactive=False)
|
1428 |
|
1429 |
with gr.Row():
|
1430 |
with gr.Column(scale=3):
|
|
|
1434 |
with gr.Row():
|
1435 |
with gr.Column():
|
1436 |
csv_generate = gr.Button(value='Generate CSV Report',
|
1437 |
+
interactive=False, variant='primary')
|
1438 |
csv_download_file = gr.File(label='Download CSV Report', visible=False)
|
1439 |
with gr.Column():
|
1440 |
html_generate = gr.Button(value='Generate HTML Report',
|
1441 |
+
interactive=False, variant='primary')
|
1442 |
html_download_file = gr.File(label='Download HTML Report', visible=False)
|
1443 |
|
1444 |
|
|
|
1553 |
|
1554 |
example_fasta.click(fn=example_fill, inputs=target_input_type, outputs=[
|
1555 |
target_id, target_gene, target_organism, target_fasta], show_progress=False)
|
1556 |
+
|
1557 |
+
|
1558 |
# example_uniprot.click(fn=example_fill, inputs=target_input_type, outputs=target_fasta, show_progress=False)
|
1559 |
# example_gene.click(fn=example_fill, inputs=target_input_type, outputs=target_fasta, show_progress=False)
|
1560 |
|
|
|
1882 |
)
|
1883 |
|
1884 |
# TODO background job from these 3 pipelines to update file_for_report
|
1885 |
+
def inquire_task(df, upload_flag):
|
1886 |
+
if upload_flag:
|
1887 |
+
if 'Y' in df.columns:
|
1888 |
+
label = 'actual CPI/CPA labels (`Y`)'
|
1889 |
+
elif 'Y^' in df.columns:
|
1890 |
+
label = 'predicted CPI/CPA labels (`Y^`)'
|
1891 |
+
else:
|
1892 |
+
return {analyze_btn: gr.Button(interactive=True),
|
1893 |
+
csv_generate: gr.Button(interactive=True),
|
1894 |
+
html_generate: gr.Button(interactive=True)}
|
1895 |
+
|
1896 |
+
return {report_task: gr.Dropdown(visible=True,
|
1897 |
+
info=f'Found {label} in your uploaded dataset. '
|
1898 |
+
'Is it compound-target interaction or binding affinity?'),
|
1899 |
+
html_report: '',
|
1900 |
+
analyze_btn: gr.Button(interactive=False),
|
1901 |
+
csv_generate: gr.Button(interactive=False),
|
1902 |
+
html_generate: gr.Button(interactive=False)}
|
1903 |
+
else:
|
1904 |
+
return {report_task: gr.Dropdown(visible=False)}
|
1905 |
+
|
1906 |
|
1907 |
file_for_report.change(fn=update_df, inputs=file_for_report, outputs=[
|
1908 |
+
html_report, raw_df, report_df, analyze_btn]).success(
|
1909 |
+
fn=lambda: [gr.Button(interactive=False)]*2 + [gr.File(visible=False)]*2 + [gr.Dropdown(visible=False)],
|
1910 |
+
outputs=[csv_generate, html_generate, csv_download_file, html_download_file, report_task]
|
1911 |
+
).then(
|
1912 |
+
fn=inquire_task, inputs=[raw_df, report_upload_flag],
|
1913 |
+
outputs=[report_task, html_report, analyze_btn, csv_generate, html_generate]
|
1914 |
+
)
|
1915 |
+
file_for_report.clear(fn=lambda: gr.Dropdown(visible=False), outputs=report_task)
|
1916 |
+
file_for_report.upload(
|
1917 |
+
fn=lambda: True, outputs=report_upload_flag
|
1918 |
+
)
|
1919 |
+
|
1920 |
+
analyze_btn.click(fn=submit_report, inputs=[raw_df, scores, filters, report_task], outputs=[
|
1921 |
+
html_report, report_df, csv_download_file, html_download_file
|
1922 |
+
]).success(fn=lambda: [gr.Button(interactive=True)] * 2,
|
1923 |
+
outputs=[csv_generate, html_generate])
|
1924 |
+
|
1925 |
+
report_task.select(fn=lambda: gr.Button(interactive=True),
|
1926 |
+
outputs=analyze_btn)
|
1927 |
+
|
1928 |
+
|
1929 |
+
def create_csv_report_file(df, file_report, progress=gr.Progress(track_tqdm=True)):
|
1930 |
try:
|
1931 |
now = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
1932 |
filename = f"reports/{Path(file_report.name).stem}_DeepSEQreen_report_{now}.csv"
|
1933 |
df.drop(labels=['Compound', 'Scaffold'], axis=1).to_csv(filename, index=False)
|
1934 |
|
1935 |
+
return gr.File(filename)
|
1936 |
except Exception as e:
|
1937 |
gr.Warning(f"Failed to generate CSV due to error: {str(e)}")
|
1938 |
+
return None
|
1939 |
+
|
1940 |
|
1941 |
+
def create_html_report_file(df, file_report, progress=gr.Progress(track_tqdm=True)):
|
1942 |
try:
|
1943 |
now = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
|
1944 |
filename = f"reports/{Path(file_report.name).stem}_DeepSEQreen_report_{now}.html"
|
1945 |
create_html_report(df, filename)
|
1946 |
+
return gr.File(filename, visible=True)
|
1947 |
except Exception as e:
|
1948 |
gr.Warning(f"Failed to generate HTML due to error: {str(e)}")
|
1949 |
+
return None
|
1950 |
+
|
1951 |
|
1952 |
html_report.change(lambda: [gr.Button(visible=True)] * 2, outputs=[csv_generate, html_generate])
|
1953 |
+
csv_generate.click(
|
1954 |
+
lambda: [gr.Button(visible=False), gr.File(visible=True)], outputs=[csv_generate, csv_download_file],
|
1955 |
+
).then(fn=create_csv_report_file, inputs=[report_df, file_for_report],
|
1956 |
+
outputs=csv_download_file, show_progress='full')
|
1957 |
+
html_generate.click(
|
1958 |
+
lambda: [gr.Button(visible=False), gr.File(visible=True)], outputs=[html_generate, html_download_file],
|
1959 |
+
).then(fn=create_html_report_file, inputs=[report_df, file_for_report],
|
1960 |
+
outputs=html_download_file, show_progress='full')
|
1961 |
|
1962 |
# screen_waiting.change(fn=check_job_status, inputs=run_state, outputs=[pair_waiting, tabs, file_for_report],
|
1963 |
# every=5)
|