Figure 5 from Inferring Drug–Gene Relationships in Cancer Using Literature-Augmented Large Language Models
Year: 2025
Type: other
Abstract: The GeneRxGPT web tool. A, Homepage. The homepage provides access to the main drug–gene inference module, as well as a link to the drug–gene network module. Users can easily test the tool using a built-in example. A comprehensive user manual, along with figure and table legends, is available on the help pages or by clicking the question mark icon in individual output sections. B, Input panel of the drug–gene inference module. Users enter a gene (gene symbol or protein name), drug [name, PubChem CID, or Simplified Molecular Input Line Entry System (SMILES) code], and cancer type into the input panel. Users can select the desired LLMs (either GPT-4o or Gemini-1.5 Pro, currently available) and choose whether the inference is based on sentence- or abstract-level retrieval. C, Output panels of the drug–gene inference module. Based on the abstracts or sentences retrieved from relevant publications in PubMed, the selected LLM generates an inference and detailed explanation of the drug–gene relationship in the specified cancer type, along with a summary of the retrieved information. References to retrieved sentences or abstracts are provided with PubMed PMIDs and hyperlinks. Additionally, detailed drug information from PubChem is presented by an informative table and the drug structure. D, Drug–gene interaction network. This module allows users to visualize a drug–target interaction network among 214 oncogenes and 144 FDA-approved drugs, either across all cancers or within a specific cancer type.
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
Open Access status: gold
OpenAlex ID: W4410344172