I'm learning to use R in data analysis. I'm getting fluent in baseR and the tidyverse, but thus far only dealt with medium throughput plate-based experiments.

I am currently trying to learn how to work with a data set of normalized RNAseq results. I have a matrix consisting of tens of thousands of rows corresponding to genes. The first two columns identify 'gene_ID' and 'gene_name'. The remaining 96 columns are different samples (different treatments, timepoints, replicates, ...).

I would like to analyze this dataset to identify differentially expressed gene based on the different conditions, perform some cluster analysis, run GESA on these clusters, and generate heatmaps at various stages of the process.

Can anybody recommend the best libraries to look into in order to achieve these goals? I started looking at several tutorials, but much of them deal with the earlier steps of RNAseq data analysis and don't really focus on these later steps once the data is already processed and cleaned up.

  • $\begingroup$ What tutorials have you read? DESeq2 starts from raw RNAseq results and goes into DE analysis, after which you can look at GSEA and similar approaches. As for data types/containers for your data based on its size, you can use plain data frames or SummarizedExperiment objects. $\endgroup$
    – Ram RS
    Jan 12, 2021 at 20:05
  • $\begingroup$ I have not yet worked myself through any specific tutorial since there seemed to be so many tools available. I started off with my 'standard approach' of turning everything into long data frames with columns for the various treatment types, but that became completely unworkable (8 columns with 5.5 million rows). So I was hoping to get a pointer about which package(s) and data formats I should start to investigate first. $\endgroup$ Jan 12, 2021 at 20:12
  • 2
    $\begingroup$ If y have normalized data, I think limma is your best bet for finding DE genes. EdgeR and DESeq are the most commonly used, but they require raw reads counts. $\endgroup$
    – swbarnes2
    Jan 12, 2021 at 20:54
  • $\begingroup$ Thanks @swbarnes2. I'll start my investigations there. $\endgroup$ Jan 13, 2021 at 0:38


Browse other questions tagged or ask your own question.