Abstract

Machine Learning is used to create predictive models by learning features from datasets. In the studies performed by Jason G. Fleischer et al. 2018 and Jana Naue et al. 2017, biomarkers are examined to predict the chronological age of humans by analysing the RNA-seq gene expression levels and DNA methylation pattern respectively. Different machine learning algorithms are used in these studies to select specific biomarkers to make age prediction. The RNA-seq gene expression (FPKM) dataset is generated using fibroblast cell lines of humans. The skin fibroblasts cells keep damage that happens with age. Epigenomic and phenotypic changes which are age-dependent are also contained in these cells. Within each individual, DNA methylation changes with age. This knowledge is used to select useful biomarkers from DNA methylation dataset. The CpGs sites with the highest correlation to age are selected as the biomarkers/features. In both these studies, specific biomarkers are analysed by machine learning algorithms to create an age prediction model.

About This Material

This is a Hands-on Tutorial from the GTN which is usable either for individual self-study, or as a teaching material in a classroom.

Questions this will address

• How to use machine learning to create predictive models from biological datasets (RNA-seq and DNA methylation)?

Learning Objectives

• Learn ageing biomarkers from RNA-seq and DNA methylation datasets
• Apply regression based machine learning algorithms
• Learn feature selection and hyperparameter optimisation