Rank protein binder designs using ipSAE_min, LIS, and pDockQ scores.
npx skills add https://github.com/junior1p/ProteinClaw --skill ipsae-junior1p
Please help me install this Agent Skill. Skill: ipsae Source: https://github.com/junior1p/ProteinClaw/tree/main/skills/ipsae Command: npx skills add https://github.com/junior1p/ProteinClaw --skill ipsae-junior1p
ipSAE provides more accurate ranking of designed protein binders than ipTM or iPAE, helping researchers prioritize candidates for experimental validation and reduce wasted lab resources. It converts predicted structures and error matrices into chain-pair scores that correlate better with binding success for designed complexes.
Rank the provided binder designs by computing ipSAE scores and return the top candidates with ipSAE_min above 0.61.
High-intent search queries and answers about installing and using this skill.
To rank protein binders from AlphaFold2 predictions, compute chain-pair ipSAE_min, LIS, and pDockQ scores using the predicted structures and PAE matrices to prioritize candidates for experimental testing.
The ipSAE score threshold for selecting designed protein binders is ipSAE_min > 0.61 for standard selection and > 0.70 for stringent selection, alongside LIS and pDockQ metric recommendations.
Yes, ipSAE works with BoltzGen and BindCraft outputs, computing chain-pair ranking scores for predicted protein complexes when PAE or equivalent error matrices are available.
Use ipSAE instead of ipTM for protein binder ranking because ipSAE provides more accurate chain-pair scores that correlate better with binding success, helping reduce wasted lab resources during experimental validation.
Yes, you need PAE or equivalent error matrices alongside predicted structures to compute ipSAE_min, LIS, and pDockQ scores for ranking designed protein binders.
jaechang-hitsRetrieve AlphaFold DB protein structures by UniProt ID and download PDB/mmCIF files.
001TMFApply hard filters and composite ranking to protein designs, outputting PASS/MARGINAL/FAIL verdicts as CSV.
adaptyvbioRank binder designs using ipSAE scores from AlphaFold2, AlphaFold3, and Boltz1 predictions.
alex-hhScore predicted protein-protein interactions using ipSAE, pDockQ, pDockQ2, and LIS metrics.
