Big data bioinformatics

Abstract

Recent technological advances allow for high throughput profiling of biological systems in a cost-efficient manner. The low cost of data generation is leading us to the "big data" era. The availability of big data provides unprecedented opportunities but also raises new challenges for data mining and analysis. In this review, we introduce key concepts in the analysis of big data, including both "machine learning" algorithms as well as "unsupervised" and "supervised" examples of each. We note packages for the R programming language that are available to perform machine learning analyses. In addition to programming based solutions, we review webservers that allow users with limited or no programming background to perform these analyses on large data compendia.

Fig. 1

The difference between supervised and unsupervised machine learning. A: In supervised machine learning, a training dataset with labeled classes, for example, case or control, is provided. A model is trained to maximally differentiate between cases and controls, and then the classes of new samples are determined. B: In an unsupervised machine learning model, all samples are unlabeled. Clustering algorithms, an example of unsupervised methods, discover groups of samples that are highly similar to each other and distinct from other samples.

Fig. 2

Unsupervised analyses discover the predominant signals in the data. For example, principle component analysis applied to a dataset combined from two large studies of breast cancer identifies the study (METABRIC or TCGA) as the most important principle component (PC1). Such confounding factors have thus far made applying unsupervised analysis methods to broad compendia challenging, so this class of methods is most frequently used within large homogenous datasets.