pheatmap2

pheatmap2 runs faster than pheatmap and supports many more distance measures. It takes the advantage of parallel distance calculation via parallelDist and fastcluster packages.

Install

Please install devtools if you haven't yet.

install.packages("devtools")

Then, you basically install the pheatmap2 package by using:

library(devtools)
install_github("altintasali/pheatmap2")

What is pheatmap?

pheatmap stands for "Pretty Heatmaps" and it definitely deserves its name. Raivo Kolde, the author of pheatmap, did an amazing job on heatmap visualization and I strongly believe that pheatmap is one of the best heatmap tools available. Dave Tang has a really blog post about how to exploit the features of pheatmap, so please check it out.

What is more on pheatmap2

pheatmap2 adds more features to pheatmap while keeping its original source code.

  1. Runs faster
    1. Distance matrix calculation: It uses parallelDist::parDist function instead of stats::dist. parDist takes the advantage of all available cores on CPU.
    2. Hierarchical clustering: It uses fastcluster::hclust function instead of stats::hclust. fastcluster::hclust is show to perform faster hierarchical clustering than stats::hclust
  2. Supports more distance measures. Here is a list of distance measures (bold text below means also supported by pheatmap):
    • correlation
    • euclidean
    • maximum
    • manhattan
    • canberra
    • binary
    • minkowski
    • geodesic
    • mahalanobis
    • fJaccard
    • bhjattacharyya
    • bray
    • chord
    • divergence
    • dtw
    • hellinger
    • kullback
    • podani
    • soergel
    • wave
    • whittaker