Training Labels

Version protein-label-v1.0 · HMM-derived weak labels

7,060,499

Labeled

4,423,272

Supervised

2,140,442

Discovery

Training Role Distribution

Supervised vs Discovery

Role Details

RoleConfidenceBinaryProteinsUsage
train_positivehigh12,784,330supervised
discovery_poolunlabeled·1,842,822discovery
background_negativemedium01,446,447supervised
auxiliary_componentmedium·320,113auxiliary
conditional_positivemedium·297,620discovery
hard_negativehigh0192,495supervised
exclude_nonprokexcluded·145,583auxiliary
review_unresolvedunlabeled·31,089auxiliary

Top Positive Seed Families

Curated Cellular Negatives