Precision Oncology: Deciphering Drug Sensitivity Through Multi-Omics Clustering and Machine Learning in Cancer Cell Lines

This study combines diverse omics datasets, including Proteomic, Transcriptomic, Genomic, Metabolomic and miRNA profiles from cancer cell lines, with Drug Sensitivity (AUC) data for 285 drugs. We used the k-means clustering algorithm to group cell lines based on multi-omics data. Drug sensitivity (A...

Full description

Saved in:
Bibliographic Details
Published in2024 International Conference on Electrical, Computer and Energy Technologies (ICECET pp. 1 - 6
Main Authors Amjad, Sara, Aziz, M. Azhar, Sufyan Beg, M. M.
Format Conference Proceeding
LanguageEnglish
Published IEEE 25.07.2024
Subjects
Accuracy
Cancer
cancer cell lines
Drugs
k-means clustering
multi-omics
Oncology
Prediction algorithms
Predictive models
Proteomics
random forest
Random forests
Sensitivity
Transcriptomics
Online AccessGet full text

Cover

More Information
Summary:This study combines diverse omics datasets, including Proteomic, Transcriptomic, Genomic, Metabolomic and miRNA profiles from cancer cell lines, with Drug Sensitivity (AUC) data for 285 drugs. We used the k-means clustering algorithm to group cell lines based on multi-omics data. Drug sensitivity (AUC) for 285 drugs were predicted using Random Forest Classifier. The resulting 5 clusters were thoroughly evaluated, suggesting drugs with over 75% accuracy for further scrutiny. Our findings indicate that the cluster-specific approach consistently outperforms the non-clustering approach. Notably, the maj ority of clusters exhibit a higher number of drugs with over 75% accuracy compared to the non-clustered approach. Moreover, the mean accuracy for identified drugs is greater for most clusters than the non-clustering approach. However, while there are some common drugs across clusters, each cluster predominantly shows a unique set of drugs. This emphasizes the importance of tailoring therapeutic strategies to the diverse molecular profiles identified within each cluster. This refined approach enhances our understanding of drug responses, supporting a cluster-centric strategy in precision oncology. It highlights the importance of customizing treatments based on the diverse molecular landscapes of cancer cell lines, contributing to the advancement of personalized treatment strategies.
DOI:10.1109/ICECET61485.2024.10698701