Multi-Omics Integration for Pancreatic Adenocarcinomas Subtyping

Cancer is a leading cause of death worldwide. Precision oncology aims to identify newmolecule-based cancer subtypes from large-scale cancer multi-omics data(۱), allowing for moreaccurate and personalized treatments. Multi-omics studies analyze high-dimensional datasets at variouslevels to reveal the complexity of cells and their environment(۲). Integrating multi-omics data hasbecome increasingly important, with machine learning playing a key role in comparing and identifyingpatterns in biological data(۳). Our study utilized multi-omics data, including transcriptomics RNAsequencing,DNA methylation, and gene mutations, to identify three molecular subtypes and assesssample similarity within the subtypes. We applied various pre-processing steps, including annotation,quality control, filtering, normalization, and feature selection. Then, we executed ten classical clusteringalgorithms to recognize patients with different molecular features using the "MOVICS" package in R.We filtered out low express genes, noncoding genes, and removed probes with detection P value > ۰.۰۱,all non-CpG probes, all SNP-related probes, all multi-hit probes, and probes located on sexchromosomes. Finally, we identified three molecular subtypes and quantified the sample similaritywithin the subtypes using the silhouette score. Our study highlights the importance of multi-omicsintegration and pre-processing steps in understanding molecular subtypes. The use of the "MOVICS"package in R provides an accessible and powerful toolset for researchers to analyze multi-omics data.Integrating multi-omics and clinical data can help identify robust and clinically actionable cancersubtypes. We hope that our findings will contribute to the development of more effective cancertherapies and personalized medicine.

کلیدواژه ها:

Multi-Omics ، Cancer Subtyping ، Pancreatic Adenocarcinoma ، MOVICS