Machine learning identifies interacting genetic variants contributing to breast cancer risk: A case study in Finnish cases and controls

Abstract

We propose an effective machine learning approach to identify group of interacting single nucleotide polymorphisms (SNPs), which contribute most to the breast cancer (BC) risk by assuming dependencies among BCAC iCOGS SNPs. We adopt a gradient tree boosting method followed by an adaptive iterative SNP search to capture complex non-linear SNP-SNP interactions and consequently, obtain group of interacting SNPs with high BC risk-predictive potential. We also propose a support vector machine formed by the identified SNPs to classify BC cases and controls. Our approach achieves mean average precision (mAP) of 72.66, 67.24 and 69.25 in discriminating BC cases and controls in KBCP, OBCS and merged KBCP-OBCS sample sets, respectively. These results are better than the mAP of 70.08, 63.61 and 66.41 obtained by using a polygenic risk score model derived from 51 known BC-associated SNPs, respectively, in KBCP, OBCS and merged KBCP-OBCS sample sets. BC subtype analysis further reveals that the 200 identified KBCP SNPs from the proposed method performs favorably in classifying estrogen receptor positive (ER+) and negative (ER−) BC cases both in KBCP and OBCS data. Further, a biological analysis of the identified SNPs reveals genes related to important BC-related mechanisms, estrogen metabolism and apoptosis.

Machine Learning Supports Long Noncoding RNAs as Expression Markers for Endometrial Carcinoma

“…Dysregulation of the expression of lincRNAs may be pervasive in human cancers and drives cancer development and progression [35]. Previous researches demonstrated that there are SNPs on LINC01376 associated with breast cancer [36,37]. Notably, it was the only lncRNA upregulated in our analysis with a logFC of 5.88, and it scored the best AUC (0.913) among all candidates.…”

“…Another lncRNA-AS that stood out in our analysis was ZNF667-AS1, with a logFC of approximately − 5.16 and an AUC of 0.717. Other studies have demonstrated that ZNF667-AS1 is commonly downregulated in several cancer types, including UCEC [36,37]. Vrba and colleagues [45] showed that it is expressed in all normal finite lifespan human cells examined to date and is downregulated or lost in immortalized human mammary epithelial cells.…”

Biological Machine Learning Combined with Campylobacter Population Genomics Reveals Virulence Gene Allelic Variants Cause Disease

“…We hypothesized that specific alleles of a single gene (i.e., porA) define the ability of Campylobacter for extraintestinal invasion and further are causative of abortion with specific alleles. This was done using a wet-lab validated data set containing 100 genomes [16,22,23] combined with machine learning using extreme gradient boosting (XGBoost) [24,25]. The ability to interrogate the predictive features emerged as a tool to determine mechanistic function in complex biological systems [26].…”

Biological machine learning combined with bacterial population genomics reveals common and rare allelic variants of genes to cause disease

“…This was done using a wet lab validated data set containing 100 genomes 6–8 using extreme gradient boosting (XGboost), which was used in biological applications previously 9 . XGboost can identify genetic variants in human GWAS as demonstrated in a Finnish study that integrated complex nonlinear interactions of SNPs 10 . The ability to interrogate the predictive features enables whiteboxing the parameters, which is emerging as a tool for deriving mechanistic function in biology 11 .…”