9e75f44f1a
Source-only snapshot of the cluster branch for WES execution. Large reference files (HPA/MANE/ensemble FASTA, model weights, ~597 MB) are omitted: they are baked into the container images at build time and mounted from the dreamdock-data PVC at runtime, and exceed the Gitea request size limit. Pipeline entry point is main.nf, which orchestrates the biotransformer, conplex and tissue modules as a single workflow. Ligand inputs are read from the eureka workspace; protein_zarr and chembl_db come from the dreamdock-data PVC.
2.4 KiB
2.4 KiB
Digital Trial Pipeline
v1 steps
| Step | Status | Model | Input | Output | Notes |
|---|---|---|---|---|---|
| 1. Metabolite Prediction | done | biotransformer | SMILES | DRUG_out.txt | intermediate step |
| 2. Drug & metabolite - screening against patient proteome | done | ConPlex | patient.fasta from digital patient step 3 (vcf2prot); "DRUG_out.txt" | "patientID_DRUG_significant_interactions.tsv" | intermediate step - pairs are sorted by scores and filtered for next step |
| 3. Biological processes/Pathway enrichment analysis | done | stringDB API | "patientID_DRUG_significant_interactions.tsv" | DRUG_patientID_network_enrichment.tsv; DRUG_patientID_network_interactions.tsv | biological processes/pathways for proteins that bind to the drug or metabolites - to use ML model to summarise output |
| 4. Drug localization |
Pipeline steps descriptions and tasks
-
Metabolite Prediction
- predicted metabolites of input drug
- filtered out intermediate metabolites and excluded from subsequent steps
-
Drug/Metabolite screen against patient proteome
- columns 2 & 3 being a pair (SMILES-ENST#) and column 4 with corresponding score for the pair in each row
- UI tasks
- allow user to input target [Ensembl ID] of test drug and we show score of interaction as main interactions; add following description:
| Score | Interaction probability | | > 0.9 | Excellent | | 0.8 - 0.9 | Acceptable | | < 0.8 | Highly unlikely |
- add a slider for adjusting the cutoff score (set default cutoff on slider to 0.75)
- show the significant interactions as table for top 20 drugs/metabolites against proteins ordered by scores below the main interaction
-
Biological processes/Pathway enrichment
- Is the drug strongly and specifically interacting with the target and/or do we have a lot of drug or metabolite off-target effects
- create ML model to create summary description using "XXX_network_enrichment.tsv"
-
Drug localization
- which organs/tissues are the interacting proteins usually localised, Human Protein Atlas tissue-enhanced protein lists for tissue proportion estimate. Heat map of proteins of interest expressed in each tissue.
- get list of protein expressed in which tissue from HPA and upload to gitlab
v2 planned steps
- Drug clearance
- mutations in Phase1/2 proteins