Clustering with R and RStudio

Clustering is one of the fundamental unsupervised machine learning algorithms. It is often used to group quantitative proteomic or RNAseq expression data to suggest sub-types of a particular cancer. This presentation covers building a distance/dissimilarity matrix, agglomerative and divisive hierarchical clustering and its associated dendrogram, and K-means clustering with a principal component and nonlinear projection of the resulting clusters. Comparing different clustering using the silhouette width is also presented. Though non-biological, the R dataset “swiss” will be used to show the basic techniques involved in clustering that can be applied to datasets of importance in cancer.