Comparison of Machine Learning Pipelines for Gene Expression Matrices

• Mateus Devino Centro Federal de Educação Tecnológica Celso Suckow da Fonseca (CEFET/RJ) https://orcid.org/0000-0002-5235-9045
• Kele Belloze Centro Federal de Educação Tecnológica Celso Suckow da Fonseca (CEFET/RJ) https://orcid.org/0000-0001-6257-2520
• Eduardo Bezerra Centro Federal de Educação Tecnológica Celso Suckow da Fonseca (CEFET/RJ) https://orcid.org/0000-0001-9177-5503

Resumo

Advances in genetic sequencing technologies have enabled the understanding of the course of diseases in a manner like never before. These technologies produce a data structure called an expression matrix, which contains gene expression values taken under certain sampling conditions. In this paper, we present preliminary work on comparing the application of different machine learning pipelines to an expression matrix. As a case study, we consider a dataset from the Gene Expression Omnibus containing gene expression levels (obtained through scRNA-seq) in the context of Breast Cancer disease. We present a generalized processing pipeline instantiation and discuss the corresponding results.

Palavras-chave: Machine learning, scRNA-seq, Breast cancer