Abstract:Predicting synergistic drug combinations can help accelerate discovery of cancer treatments, particularly therapies personalized to a patient's specific tumor via biopsied cells. In this paper, we propose a novel setting and models for in-context drug synergy learning. We are given a small "personalized dataset" of 10-20 drug synergy relationships in the context of specific cancer cell targets. Our goal is to predict additional drug synergy relationships in that context. Inspired by recent work that pre-trains a GPT language model (LM) to "in-context learn" common function classes, we devise novel pre-training schemes that enable a GPT model to in-context learn "drug synergy functions". Our model -- which does not use any textual corpora, molecular fingerprints, protein interaction or any other domain-specific knowledge -- is able to achieve competitive results. We further integrate our in-context approach with a genetic algorithm to optimize model prompts and select synergy candidates to test after conducting a patient biopsy. Finally, we explore a novel task of inverse drug design which can potentially enable the design of drugs that synergize specifically to target a given patient's "personalized dataset". Our findings can potentially have an important impact on precision cancer medicine, and also raise intriguing questions on non-textual pre-training for LMs.

What problem does this paper attempt to address?

### Problems the Paper Attempts to Solve This paper aims to address several key challenges in personalized drug synergy prediction and drug design: 1. **Personalized Drug Synergy Prediction**: - In cancer treatment, predicting drug synergy for a specific patient's tumor through biopsy cells can accelerate the discovery of new treatments. However, due to the extensive mutations and different genetic backgrounds of tumor cells in different patients, predicting these synergies is very challenging. - The paper proposes a new method that uses a small-scale "personalized dataset" (usually containing 10-20 drug synergy relationships) to predict more drug synergy relationships in the context of specific cancer cell targets. 2. **Few-Shot Learning**: - Due to the exponential growth in the number of drug combinations, testing the synergy or antagonism of a large number of drug pairs under laboratory conditions is very expensive. Therefore, the paper explores how to predict drug synergy in a few-shot setting, where only a small number of training examples are available. - The paper proposes a context learning method based on the GPT model, which can generalize from a small number of known drug synergy relationships to unseen drugs and patient cell lines without external knowledge. 3. **Inverse Drug Design**: - The paper also proposes a new task—Inverse Synergistic Drug Structure Design (ISDSD), which involves generating or retrieving new molecules that have synergistic effects with existing drugs in the context of a specific patient. - This method may provide new avenues for the discovery of personalized drug candidates in the future. ### Main Contributions 1. **SynerGPT Model**: - The paper proposes the SynerGPT model, which predicts drug synergy in a few-shot setting through context learning without any external knowledge (such as protein interaction networks or patient cell line characteristics). - The model performs well on personalized datasets and can generalize to unseen drugs and cell lines. 2. **Context Optimization**: - The paper proposes a context optimization method based on a genetic algorithm to select the best prompts to improve the model's performance in predicting the synergy of unknown drugs or cell lines. - This optimization method helps develop standardized drug synergy detection methods. 3. **Inverse Drug Design**: - The paper explores how to use the GPT model to generate or retrieve new molecules that have synergistic effects with existing drugs in the context of a specific patient. - This method not only helps discover new drug candidates but also improves the model's interpretability by "visualizing" SynerGPT's understanding of unknown drugs. ### Experimental Results - **Performance of the BERT Model**: - The paper's experiments show that even without using external knowledge, the BERT model outperforms existing knowledge-enhanced models (such as GraphSynergy) in the drug synergy prediction task. - When using random labels instead of drug and cell names, the model's performance does not significantly decline, suggesting that the transformer architecture itself may be the key factor explaining BERT's performance. - **Few-Shot and Zero-Shot Learning**: - In few-shot and zero-shot settings, the SynerGPT model outperforms existing baseline models, especially in predicting the synergy of unknown drugs or cell lines. - The context optimization method significantly improves the model's performance in few-shot settings. - **Inverse Drug Design**: - Preliminary experimental results show that the GPT model can generate new molecules that have synergistic effects with existing drugs, providing new ideas for personalized drug design. Overall, this paper introduces the SynerGPT model and context optimization methods, providing new solutions for personalized drug synergy prediction and inverse drug design, with significant clinical application potential.

SynerGPT: In-Context Learning for Personalized Drug Synergy Prediction and Drug Design

CancerGPT: Few-shot Drug Pair Synergy Prediction using Large Pre-trained Language Models

A Deep Neural Network for Predicting Synergistic Drug Combinations on Cancer

From cell lines to cancer patients: personalized drug synergy prediction

DeepSynergy: predicting anti-cancer drug synergy with Deep Learning

Drug Synergistic Combinations Predictions via Large-Scale Pre-Training and Graph Structure Learning

CancerGPT for few shot drug pair synergy prediction using large pretrained language models

SynerGNet: A Graph Neural Network Model to Predict Anticancer Drug Synergy

How to Predict Effective Drug Combinations - Moving beyond Synergy Scores

Unlocking the therapeutic potential of drug combinations through synergy prediction using graph transformer networks

GraphSynergy: a network-inspired deep learning model for anticancer drug combination prediction

Anti-cancer Drug Synergy Prediction in Understudied Tissues using Transfer Learning

DTSyn: a Dual-Transformer-based Neural Network to Predict Synergistic Drug Combinations

Multi-way relation-enhanced hypergraph representation learning for anti-cancer drug synergy prediction

RECOVER: sequential model optimization platform for combination drug repurposing identifies novel synergistic compounds in vitro

Deep learning-based multi-drug synergy prediction model for individually tailored anti-cancer therapies

An In Silico Method for Predicting Drug Synergy Based on Multitask Learning

SynergyX: a multi-modality mutual attention network for interpretable drug synergy prediction

MMSyn: A New Multimodal Deep Learning Framework for Enhanced Prediction of Synergistic Drug Combinations

CCSynergy: an integrative deep-learning framework enabling context-aware prediction of anti-cancer drug synergy

Interpretable Drug Synergy Prediction with Graph Neural Networks for Human-AI Collaboration in Healthcare