跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2610.07997 · v2

Can phenotypic activity be predicted without experimental readouts?

首次提交 2026-10-06 16:56
版本更新 2026-11-17 01:25
雷达观测 2026-11-17 16:10
arXiv ↗PDF ↗

摘要

Molecular encoders contrastively pretrained on paired molecule-morphology data, such as CLOOME and CellCLIP, have been proposed as cheap surrogates for phenotypic prediction, avoiding the need to run a Cell Painting assay. We evaluate this idea under a protocol designed to control for two confounds that can inflate apparent performance: leakage across an encoder's own pretraining boundary, and the correlation between phenotypic activity and simple cytotoxicity. Testing six representations, including a non-pretrained MLP control matching CLOOME's input and layer count, on two Cell Painting screens of different scale and origin, we find that once these confounds are controlled for, CLOOME and CellCLIP show no clear advantage over plain physicochemical descriptors, also under stricter activity definitions and with non-linear probes, and that toxicity is generally easier to predict than genuine phenotypic activity across representations. The protocol is not tied to these two models, and we suggest that leakage-aware, confound-controlled evaluation should be standard practice before phenotype-pretrained encoders are trusted as surrogates for predictive phenotypic drug discovery.

作者备注:Accepted to the ML4Molecules: Agentic Systems for Molecular Sciences Workshop at the 40th Conference on Neural Information Processing Systems (NeurIPS 2026)

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

本站首次收录为 v2,暂无更早版本可对照。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v2 · 2026-11-17 01:25 · 最新观测版本

Can phenotypic activity be predicted without experimental readouts?

Molecular encoders contrastively pretrained on paired molecule-morphology data, such as CLOOME and CellCLIP, have been proposed as cheap surrogates for phenotypic prediction, avoiding the need to run a Cell Painting assay. We evaluate this idea under a protocol designed to control for two confounds that can inflate apparent performance: leakage across an encoder's own pretraining boundary, and the correlation between phenotypic activity and simple cytotoxicity. Testing six representations, including a non-pretrained MLP control matching CLOOME's input and layer count, on two Cell Painting screens of different scale and origin, we find that once these confounds are controlled for, CLOOME and CellCLIP show no clear advantage over plain physicochemical descriptors, also under stricter activity definitions and with non-linear probes, and that toxicity is generally easier to predict than genuine phenotypic activity across representations. The protocol is not tied to these two models, and we suggest that leakage-aware, confound-controlled evaluation should be standard practice before phenotype-pretrained encoders are trusted as surrogates for predictive phenotypic drug discovery.

查看此版本解读 →查看此版本原文 ↗