is a Data Source.
SEP-L1000 (Side Effect Prediction based on L1000 data) is a web portal from the Ma'ayan Laboratory that presents adverse drug reaction (ADR) predictions for 20,412 small molecules profiled by the LINCS L1000 project across 1,054 side effects. Predictions come from multi-label extremely randomized trees classifiers trained on known ADRs from SIDER and OFFSIDES, using L1000 gene expression signatures, Gene Ontology enrichment vectors, chemical fingerprints, and cell morphology profiles as features. The processed feature and label matrices are available for download.
Warning: No license entered
No external IDs available
| ID | Name | URL | Category | Format | Description |
|---|---|---|---|---|---|
| sep-l1000.portal | SEP-L1000 Web Portal | SEP-L1000 | Graphical Interface | http | Web portal for browsing and searching... |
| sep-l1000.faers-offsides | SEP-L1000 FAERS OFFSIDES Side Effect Matrix | FAERS_offsides_PTs.csv.gz (452.2 KB) | Product | csv | Off-label drug-ADR matrix (684 drugs ... |
| sep-l1000.sider | SEP-L1000 SIDER Side Effect Matrix | SIDER_PTs.csv.gz (165.9 KB) | Product | csv | On-label drug-ADR matrix (834 drugs b... |
| sep-l1000.lincs-signatures | SEP-L1000 LINCS L1000 CD Signatures | LINCS_Gene_Experssion_signatures_CD.csv.gz (38.6 MB) | Product | csv | Characteristic direction gene express... |
| sep-l1000.go-signatures | SEP-L1000 GO-Transformed Signatures | GO_transformed_signatures_PAEA.csv.gz (219.9 MB) | Product | csv | Gene Ontology transformed L1000 signa... |
| sep-l1000.maccs-fingerprints | SEP-L1000 MACCS Fingerprints | MACCS_bitmatrix.csv.gz (1.2 MB) | Product | csv | 166-bit MACCS chemical fingerprint ma... |
| sep-l1000.mlpcn-morphology | SEP-L1000 MLPCN Morphological Profiles | MLPCN_morplological_profiles.csv.gz (52.9 MB) | Product | csv | Cell morphology profiles (812 Cell Pa... |
| sep-l1000.strongest-signatures-metadata | SEP-L1000 Strongest Signature Metadata | meta_LINCS_strongest_signatures.csv (31.8 MB) | Product | csv | Metadata mapping each LINCS L1000 com... |
| sep-l1000.probes-metadata | SEP-L1000 Probe Metadata | meta_Probes_info.csv (361.7 KB) | Product | csv | Mapping of L1000 probes to gene symbols. |
| sep-l1000.smiles-metadata | SEP-L1000 Compound SMILES Metadata | meta_SMILES.csv (3.8 MB) | Product | csv | Compound metadata for 41,774 LINCS L1... |
| ID | Name | URL | Category | Format | Relation | Description |
|---|---|---|---|---|---|---|
| drugmonizome.libraries ⚠ | Drugmonizome Drug Set Libraries | libraries | Product | json | was derived from | The 35 Drugmonizome drug set librarie... |
SEP-L1000 is a web portal from the Ma’ayan Laboratory at the Icahn School of Medicine at Mount Sinai that prioritizes adverse drug reactions (ADRs) for approved drugs and preclinical small molecules. It was built with support from the BD2K-LINCS Data Coordination and Integration Center and the NIH Illuminating the Druggable Genome program, and accompanies Wang, Clark and Ma’ayan 2016 (Bioinformatics).
Known on-label ADRs come from SIDER and off-label ADRs from OFFSIDES (processed FAERS reports), with ADR terms mapped to MedDRA preferred terms. Features for each compound include characteristic direction signatures from the strongest LINCS L1000 experiment, Gene Ontology enrichment vectors derived from those signatures, MACCS and PubChem chemical fingerprints, and cell morphology profiles from the Broad Institute MLPCN profiling dataset. DrugBank names and synonyms were used to map marketed drugs to L1000 compounds. Extremely randomized trees classifiers, one per ADR, were applied to all profiled compounds; the portal reports predictions for 20,412 compounds and 1,054 side effects.
The portal lets users browse drugs and side effects, search, and view side effect-GO term and side effect-phenotype networks. The download page provides the processed side effect matrices, signature and fingerprint matrices, morphology profiles, and metadata as CSV files. The site is served over HTTP only; the HTTPS address does not respond. It was online and its database backend was answering queries on 2026-10-19.
No license is stated on the site or in the GitHub repository. Users are asked to cite Wang et al. 2016.
Created: October 19, 2026 | Last modified: October 19, 2026