Bioinformatics Seminar
Time:
Venue: Na
6 October 2020
NaBreast cancer subtype classification at single-cell level
Goknur GinerWEHI Bioinformatics
As the most commonly diagnosed cancer ; breast cancer imposes a large health burden on Australians. While there are a number of breast cancer treatment options available ; improving the effectiveness of these treatments is one of the major clinical challenges. One of the biggest impediments to a patient's outcome is the inaccurate subtype classification of their tumour due to the molecular and cellular heterogeneity ; i.e. the diverse collections of molecular signatures e.g. genes ; present within that tumour. Despite their substantial contribution to disease prognosis ; the current approaches to classify breast tumors overlook the critical concept of heterogeneity within a tumour. This concern is now surmountable using single-cell RNA sequencing (scRNA-seq) technology. This talk will focus on improving breast tumour subtype classifications using scRNA-seq ; which can identify the heterogeneity within individual tumours. To test this concept ; we performed analyses on Fluidigm C1 sc-RNA-seq of primary breast cancer data set from GSE75688 with 515 cells from 11 patients with distinct molecular subtypes representing the four main subtypes of breast cancer patients: luminal A; luminal B; HER2; and triple negative breast cancer. We selected best performing classification algorithms for RNA-seq data ; including qtQDA ; an RNA-seq classifier ; support vector machines (SVM) and k nearest neighbourhood (kNN) machine learning approaches and glmnet ; a general regression approach to classify each cell within individual tumors. Our preliminary results showed that by classifying breast tumours at single-cell level rather than treating them as a bulk cell population ; we can provide more precise proportions of molecular heterogenity within individual tumors. As a result of this ; we predict that more appropriate treatment regimens ; such as chemotherapy ; hormone therapy and immunotherapy can be administered to those patients most likely to benefit from them.;