An Improved, Assay Platform Agnostic, Absolute Single Sample Breast Cancer Subtype Classifier

Abstract

While intrinsic molecular subtypes provide important biological classification of breast cancer, the subtype assignment of individuals is influenced by assay technology and study cohort composition. We sought to develop a platform-independent absolute single-sample subtype classifier based on a minimal number of genes. Pairwise ratios for subtype-specific differentially expressed genes from un-normalized expression data from 432 breast cancer (BC) samples of The Cancer Genome Atlas (TCGA) were used as inputs for machine learning. The subtype classifier with the fewest number of genes and maximal classification power was selected during cross-validation. The final model was evaluated on 5816 samples from 10 independent studies profiled with four different assay platforms. Upon cross-validation within the TCGA cohort, a random forest classifier (MiniABS) with 11 genes achieved the best accuracy of 88.2%. Applying MiniABS to five validation sets of RNA-seq and microarray data showed an average accuracy of 85.15% (vs. 77.72% for Absolute Intrinsic Molecular Subtype (AIMS)). Only MiniABS could be applied to five low-throughput datasets, showing an average accuracy of 87.93%. The MiniABS can absolutely subtype BC using the raw expression levels of only 11 genes, regardless of assay platform, with higher accuracy than existing methods.

Tumor microenvironment-aware, single-transcriptome prediction of microsatellite instability in colorectal cancer using meta-analysis

“…Attempts to create an absolute predictor for subtype classification of cancer and stratification of patients by applying relationships or ratios between two genes, not the expression value of the gene itself, are ongoing 22 , 23 . MAP is an absolute classifier, not relative, and was developed to reflect tumor molecular characteristics, immune-related signatures, and tumor-infiltrating immune cells in TME of CRC.…”

Breast Cancer Subtypes Classification with Hybrid Machine Learning Model

“…Beykikhoshk et al 54 applied deep learning architecture to classify gene expression signature of breast cancer subtypes namely luminal A and luminal B and calculated the individual patient biomarkers scores. A single sample platform independent subtype classifier with minimal number of genes that yield high classification accuracy by applying random forest classification algorithm was reported in Seo et al 55 ML models were also employed to explore the interaction mechanisms of genes in identifying the five inherent categories of breast cancer with RNA-Seq data. 56 Xie et al 57 investigated the performance of MR multiparametric radiomics in differentiating the breast cancer subtypes with several ML models.…”

Tumor microenvironment-aware, single-transcriptome prediction of microsatellite instability in colorectal cancer using meta-analysis

“…Attempts to create an absolute predictor for subtype classification of cancer and stratification of patients by applying relationships or ratios between two genes, not the expression value of the gene itself, are ongoing 19,20 . MAP is an absolute classifier, not relative, and was developed to reflect tumor molecular characteristics, immune-related signatures, and tumor-infiltrating immune cells in TME of CRC.…”