Skip to content
Controlled benchmarkCPJUMP1 profilesOpen-source researchModel evaluation pending

Search perturbation-induced morphology with biological language

Perturb-LM is currently a text-to-morphology benchmark. It validates leakage-aware retrieval over Cell Painting profiles while keeping image-level retrieval as a longer-term direction.

Research question

Can frozen biomedical language representations retrieve perturbation-induced cellular morphology better than strong identifier-stripped lexical controls?
4,524
profiles
904
morphology features
641
full benchmark queries
0.2513
identifier-stripped TF-IDF mAP

mAP 95% query-bootstrap CI: 0.2445 to 0.2582. This is a lexical control, not a learned-model result.

Illustrative interface demo — not real model output

Why leakage matters

Treatment names, target sequences, plates, wells, and replicates can make retrieval look stronger than it is. Perturb-LM treats those shortcuts as measurable failure modes.

Naive evaluation

Can reward exact treatment identifiers, target sequences, replicate structure, and acquisition artifacts.

Leakage-aware evaluation

Removes identifiers, reports evaluable queries, and tests held-out and plate/well-aware conditions.

How the system works

1

Biological language query

Researcher describes a phenotype or mechanism.

2

Frozen text representation

The real biomedical encoder evaluation is pending.

3

Lightweight alignment

A small projection is the next controlled experiment.

4

Morphology-profile retrieval

Profile-space ranking before any image-level claim.

5

Perturbation-level evaluation

Metrics aggregate at the perturbation level.

Validated foundation

QC, leakage checks, deterministic queries, train-only preprocessing, and reproducible smoke workflows are in place.

Lexical benchmark

Identifier-stripped TF-IDF establishes the primary lexical control that future models must beat.

Pending learned model

Frozen text embeddings and a lightweight projection are the next experiment, not a completed result.

Current limitations

  • The real biomedical text-alignment experiment has not started.
  • Split-specific learned-model results remain pending.
  • Held-out-batch evaluation is unavailable.
  • The additional profile plate changes the feature schema and is only a compatibility investigation.
  • The current benchmark is text-to-morphology-profile retrieval, not validated text-to-image retrieval.

Roadmap

Frozen embedding baseline
Regularized linear projection
Replicate consensus
Hard negatives
Second-batch harmonization
Image linkage
Interpretable morphology attribution