Handing over disease/target evidence generation from Uniprot

[DSuveges]

As the Uniprot team cannot support/run their evidence generation pipeline, we take over from them.

• The code is here: open-targets-core-db

Tasks:

• [x] Run evidence generation
• [x] Uniprot team handing over ownership over the repo.
• [ ] Based on analytics on the evidence dataset make a decision on the future of the pipeline (options: keep java as it is, re-implement in Python, drop the entire dataset if not bringing enough value)

[DSuveges]

Running evidence genration

Had to set the following things:

• One of the Java requirement's source was missing.
• maven needed to be installed
• The shell scripts needed to be updated so urls are pointing to ols4 APIs.
• Python has to point to Python3

Once all updates were in place, the evidence generation took 1h 57m 50s, the resulting dataset contains 38,485 evidence string, which is a modest (109) increse compared to the most recent data update (28-04-2023) Once we own the repo, we can commit these updates to make the run smoother.

[DSuveges]

Owning the code-base

The original codebase is forked to opentargets organization (https://github.com/opentargets/uniprot-evidence-parser). With some changes including:

• POM file contains dependency repository so now compiles.
• OLS API URLs are updated in the bash scripts.
• License (Apache v2.0) added to the repo.

[DSuveges]

• The handover has happened.
• New evidence set uploaded to: gs://otar011-uniprot/cttv011-02-01-2024.json.gz
• Evidence set validated.
• Number of evidence: 38,812 from 38,376. Modest increase with 436 evidence.