I want to prepare a matrix of gene expression to analyse TCGA LAML data.
The required data is available at TCGA LAML - Gene expression quantification.
The following is an example of the kind of data retrieved from the TCGA data portal:
ENSG00000000003.13 9 ENSG00000000005.5 1 ENSG00000000419.11 661 ENSG00000000457.12 1434 ENSG00000000460.15 1211 ENSG00000000938.11 405 ENSG00000000971.14 251 ENSG00000001036.12 786 ENSG00000001084.9 4423 ENSG00000001167.13 5038 ENSG00000001460.16 246 ENSG00000001461.15 2773 ENSG00000001497.15 1932 ENSG00000001561.6 1693 ENSG00000001617.10 55 ENSG00000001626.13 8 ENSG00000001629.8 4629
Each row displays, for each gene, the quantification of read counts by the htseq tool.
My question is, how do I have to preprocess the data (e.g. regarding gene length) before using data for subsequent analysis?
In this case I will not consider normalisation (such as loess) and batch effects removal (e.g. with combat).