nf-core / nf-core/proteinfold

Benchmarking prediction accuracy of different modes against input experimental structure.

Open
#340 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement feature-request
Dominant language
HTML
Stars
115
Forks
75
PR merge metrics
No merged PRs in 30d

Description

Description of feature

One use case for running multiple modes in parallel is to benchmark the accuracy of different methods when there is a known experimental structure for the prediction target.

It could be a nice option to enable users to input an experimental structure and score how well the predictions from each of the modes matches the provided experimental structure.

The boltz authors used OpenStructure to judge the quality of predictions for general molecules which would likely be an appropriate strategy here.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing how the workflow runs multiple prediction modes and how users could provide an experimental structure. Investigate OpenStructure as the proposed scoring approach. Done should include a documented way to compare each mode's predictions against the provided structure, with accuracy results that identify the best-performing mode.

Written by the indexing model from the issue text.

Assessment

Domain
machine-learning, testing-qa
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
25/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.