Back

SHOT-CCR: Biologically guided adversarial training for test-time adaptation in cellular morphology

Dee, W.; Wenteler, A.; Seal, S.; Morris, O.; Slabaugh, G.

2026-04-02 cell biology
10.64898/2026.03.31.715531 bioRxiv
Show abstract

AO_SCPLOWBSTRACTC_SCPLOWPervasive batch effects are a common issue, especially in recent large-scale Cell Painting datasets, which have been produced to aid AI-enhanced drug discovery efforts. Technical differences arising from experiments carried out in different batches can cause models to fail to generalize to unseen batches, despite good predictive performance "within batch". We propose a biologically grounded test-time adaptation framework, SHOT-CCR, which uses cell-invariant gradient reversal to decouple morphological signal from experimental confounders. Our approach performs 4.5% better than the current RxRx1 benchmark, classifying 1,139 classes of siRNA genetic perturbations with 91.6% accuracy. We deliver consistent results over four distinct cell types and two prominent Cell Painting datasets - RxRx1 and a subset of JUMP-CP. Across 484 classes of CRISPR perturbations in JUMP-CP our method improves accuracy by 15.7%.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

1
Cell Systems
167 papers in training set
Top 0.4%
18.5%
2
Nature Machine Intelligence
61 papers in training set
Top 0.1%
18.5%
3
Nature Methods
336 papers in training set
Top 1%
9.1%
4
Nature
575 papers in training set
Top 4%
6.3%
50% of probability mass above
5
Nature Communications
4913 papers in training set
Top 29%
6.3%
6
Nature Medicine
117 papers in training set
Top 0.4%
4.8%
7
Science
429 papers in training set
Top 8%
3.9%
8
Proceedings of the National Academy of Sciences
2130 papers in training set
Top 26%
2.3%
9
Nature Cell Biology
99 papers in training set
Top 2%
2.1%
10
Nature Genetics
240 papers in training set
Top 4%
1.9%
11
Cell Reports
1338 papers in training set
Top 22%
1.9%
12
Journal of Cell Biology
333 papers in training set
Top 2%
1.9%
13
Nature Biotechnology
147 papers in training set
Top 4%
1.9%
14
PLOS Computational Biology
1633 papers in training set
Top 17%
1.7%
15
PLOS ONE
4510 papers in training set
Top 57%
1.5%
16
Genome Biology
555 papers in training set
Top 6%
1.1%
17
eLife
5422 papers in training set
Top 52%
0.9%
18
Nature Biomedical Engineering
42 papers in training set
Top 1%
0.9%
19
Communications Biology
886 papers in training set
Top 17%
0.9%
20
Scientific Reports
3102 papers in training set
Top 71%
0.9%
21
Nucleic Acids Research
1128 papers in training set
Top 16%
0.9%
22
iScience
1063 papers in training set
Top 27%
0.9%
23
Patterns
70 papers in training set
Top 3%
0.6%
24
Science Advances
1098 papers in training set
Top 33%
0.6%