| bgc-accession model is a Named Entity Recognition (NER) model that identifies and annotates the accession number of biosynthetic gene clusters in texts. | |
| The model is a fine-tuned BioBERT model and the training dataset is available in https://gitlab.com/maaly7/emerald_bgcs_annotations | |
| Testing examples: | |
| 1. The genome sequences of Leptolyngbya sp. PCC 7375 (ALVN00000000) and G. sunshinyii YC6258 (NZ_CP007142.1) were obtained previously.36,59 | |
| 2. K311 was sequenced (NCBI accession number: JN852959) and analyzed with FramePlot and 18 genes were predicted to be involved in echinomycin biosynthesis (Figure 2). | |
| 3. The mar cluster was sequenced and annotated and the complete sequence was deposited into Genbank (accession KF711829). |