Ontology-Guided Pathway Activity Identifies a Cell-Intrinsic Defense Response Program Associated with MEK Inhibitor Sensitivity

Authors

Ndubuisi, C. W.

Abstract

Predicting cancer drug response from gene expression requires models that expose which biological pathways drive cell-line-specific sensitivity. We introduce Gene-Ontology Pathway Attention (GOPA), whose core module computes deterministic attention weights softmax(x_c * B_hat) from expression and the column-normalized gene-term annotation matrix with no learned parameters. Applying GOPA to 542 drugs across the GDSC panel under leave-cell-line-out evaluation, we find that defense response pathways predict sensitivity to kinase inhibitors in GDSC, with the strongest and most confound-resistant signal in MEK/MAPK inhibitors. This association shows cross-assay support in PRISM (11 of 11 overlapping drugs; all p < 0.002) and retains 72% signal strength after controlling for five confounds, but was not reproduced in the gCSI panel, which used a different response metric, a smaller sample, and lacked three of the four GDSC MEK inhibitors. This indicates that the finding may be MAPK-pathway-specific and assay-dependent. On the 16-drug benchmark, XGBoost achieves the lowest RMSE (1.185), while GOPA (1.226) is the strongest neural model. On the full 542-drug panel, target-encoded XGBoost matches GOPA on RMSE (1.322 vs. 1.327). GOPA achieves higher residual Pearson correlation (0.469; 95% CI [0.453, 0.485]) than target-encoded XGBoost (0.391; 95% CI [0.379, 0.404]), with a paired difference of +0.078 [0.062, 0.093]. GOPA wins on 363 of 539 drugs (67.3%; Wilcoxon p = 1.1 x 10^-23). When pathway representations are evaluated with matched downstream learners, simple gene-set projections achieve equivalent prediction, indicating that GOPA's value lies in its deterministic, population-comparable pathway summaries rather than representational superiority. A controlled geometry comparison shows that Poincare-ball embeddings preserve GO graph distances better than Euclidean embeddings (rho = 0.732 vs. 0.474), while Euclidean embeddings achieve stronger ancestor retrieval. Neither geometry improves prediction (Delta RMSE = +0.008; p = 0.18).

Preprint server: bioRxiv
The authors list and abstract were imported from bioRxiv on 30 Jul 2026.

Stats

1-terrible, 9-excellent. How would you rate this preprint? Sign in in to submit your rating.

Post a comment

You need to be signed in to post comments. You can sign in here.

Comments

There are no comments yet.