Predicting Therapeutic Outcome via Aligning Patient-Specific Knowledge Graph and Gene-Level Perturbation Representations

2026-07-05

AI Summary

Accurately predicting individual drug response from pre-treatment transcriptomes remains challenging due to the scarcity of clinical response labels and post-treatment molecular data. This work proposes a novel framework that first constructs patient-specific gene regulatory networks integrated with drug target information, then leverages a LINCS L1000–pretrained gene attention model to simulate drug perturbation effects. A CLIP-style contrastive learning strategy aligns personalized knowledge graphs with perturbation representations in a shared latent space. The approach uniquely unifies mechanistic interpretability with dynamic perturbation modeling and enables zero-shot transfer. It consistently outperforms state-of-the-art methods across multiple TCGA partitioning schemes and achieves a 5.6% improvement in zero-shot AUROC on the I-SPY2 trial, while yielding stable, biologically interpretable attributions at the gene and pathway levels.

Abstract

Accurate prediction of patient-specific therapeutic response from pre-treatment transcriptomes is hindered by the scarcity of matched clinical response labels and post-treatment molecular profiles. Preclinical transfer-learning models can simulate drug-induced expression changes but are often hard to interpret and unstable, whereas knowledge-graph methods provide mechanistic context yet remain static and fail to capture drug-induced transcriptomic perturbation dynamics. We propose PREDIKTOR, a patient-centered multi-view framework that aligns a personalized network view with a transferable transcriptomic perturbation view to predict clinical drug response. For each patient, we construct an individualized gene regulatory network from tumor expression using DysRegNet and augment it with drug-target links from DrugBank; a graph neural encoder yields a drug-centric, mechanistically grounded embedding. In parallel, a frozen condition-specific gene-gene attention model pretrained on LINCS L1000 generates a simulated post-perturbation transcriptomic profile for the same patient-drug pair. We align the two views in a shared latent space via a CLIP-style contrastive objective with drug-context hard negatives, then concatenate the representations for end-to-end response classification. On TCGA, PREDIKTOR consistently outperforms state-of-the-art baselines under patient-, drug-, and tissue-split evaluations, and transfers zero-shot to the I-SPY2 trial, improving AUROC by 5.6% over competing methods. The aligned embeddings yield stable gene and pathway attributions that recover known mechanisms, supporting actionable and interpretable precision oncology.

Problem

• therapeutic response prediction
• patient-specific
• transcriptomic perturbation
• knowledge graph
• drug response

Innovation

• knowledge graph alignment
• transcriptomic perturbation
• contrastive learning
• zero-shot transfer
• interpretable precision oncology