Biol Direct. 2025 Jan 21;20(1):11. doi: 10.1186/s13062-025-00598-y.
ABSTRACT
Advances in sequencing technologies are reshaping clinical diagnostics, prompting the development of new software tools to decipher big data. To this end, we developed functional genomic imaging (FGI), a visualization tool designed to assist clinicians in interpreting RNA-Seq results from patient samples. FGI uses weighted gene co-expression network analysis (WGCNA), followed by a modified Phenograph clustering algorithm to identify co-expression gene clusters. These gene modules were annotated and projected onto a t-SNE map for visualization. Annotation of FGI gene clusters revealed three categories: tissue-specific, functional, and positional. These clusters may be used to build tumor subtypes with pre-annotated functions. At the multi-cancer cohort level, tissue-specific clusters are enriched, whereas at the single cancer level, such as in lung cancer or ovarian cancer, positional clusters can be more prominent. Moreover, FGI analysis could also reveal molecular tumor subtypes not documented in clinical records and generated a more detailed co-expression gene cluster map. Based on different levels of FGI modeling, each individual tumor sample can be customized to display various types of information such as tissue origin, molecular subtypes, immune activation status, stromal signaling pathways, cell cycle activity, and potential amplicon regions which can aid in diagnosis and guide treatment decisions. Our results highlight the potential of FGI as a robust visualization tool for personalized medicine in molecular diagnosis.
PMID:39838379 | DOI:10.1186/s13062-025-00598-y