Novel progressive deep learning algorithm for uncovering multiple single nucleotide polymorphism interactions to predict paclitaxel clearance in patients with nonsmall cell lung cancer

Wei Chen ^1,2, Haiyan Zhou ², Mingyu Zhang ²

⁰School of Basic Medicine and Clinical Pharmacy China Pharmaceutical University Nanjing China.

Cancer Innovation +

|

July 1, 2024

View abstract on PubMed

Summary

A new machine learning algorithm, GEP-CSIs, analyzes genetic variations (SNPs) impacting paclitaxel drug clearance. This method identifies complex gene interactions to personalize cancer chemotherapy for better effectiveness.

Area Of Science

Pharmacogenomics
Bioinformatics
Computational Biology

Background

Paclitaxel clearance varies significantly among individuals, impacting chemotherapy effectiveness.
Genetic polymorphisms, specifically multiple single nucleotide polymorphisms (SNPs), are key drivers of this metabolic variability.
Existing bioinformatics methods are insufficient for analyzing complex SNP interactions in drug metabolism.

Purpose Of The Study

To develop a novel, efficient algorithm for analyzing interactions among multiple single nucleotide polymorphisms (SNPs).
To investigate the influence of SNP combinations on paclitaxel clearance in non-small cell lung cancer patients.
To enhance the understanding of genetic factors affecting drug metabolism and personalize chemotherapy.

Main Methods

Developed the GEP-CSIs data mining algorithm, an advanced version of GEP utilizing linear algebra for discrete variables.
Applied the GEP-CSI algorithm to paclitaxel clearance data and genetic polymorphisms in non-small cell lung cancer patients.
Utilized a primary and validation dataset split for robust analysis and validation of identified SNP combinations.

Main Results

Identified and validated 1184 three-SNP combinations with the highest fitness scores, indicating significant impact on paclitaxel clearance.
Discovered indirect influence of SNPs in SERPINA1, ATF3, and EGF genes on paclitaxel clearance by coordinating other significant genes.
Found a novel interaction among SNPs in FLT1, EGF, and MUC16 genes, with their proteins confirmed to interact in a protein-protein interaction network.

Conclusions

Successfully developed an effective deep-learning algorithm (GEP-CSIs) for mining complex SNP interactions.
The algorithm leverages paclitaxel clearance data and individual genetic polymorphisms for nuanced analysis.
This approach holds promise for improving the understanding and personalization of cancer chemotherapy.

Keywords:

deep learning algorithm paclitaxel single nucleotide polymorphisms