search

USEPA / CompTox-ToxCast-tcplFit2

A Concentration-Response Modeling Utility
https://usepa.github.io/CompTox-ToxCast-tcplFit2/
Other
1 stars 0 forks source link

88 flipped hitc #89

Closed brown-jason closed 2 months ago

brown-jason commented 3 months ago

Cutoff now takes into sign(top) into account.

madison-feshuk commented 3 months ago

See https://github.com/USEPA/ToxCast-Internal/tree/main/analysis/tcplfit2_hitc_toplikelihood for a standalone analysis reviewing this issue and fix (Rmd and html). Slides featuring the reprocessing results on the ToxCast database, which were presented to HTTr/HTPP/ToxCast teams on 6/10, are also provided.

ToxCast team approves this toplikelihood fix for its invitrodb v4.2 release. In the release note, an known issue will be listed related to poly2 reparameterization which may be skewing toplikelihood calculation #90