Hi, First, I would like to thank the GLASS team to share this great resource. We are planning to use GLASS RNA-seq data to examine the differential expression of several genes of interest between paired primary and recurrent GBM tumors. I wonder what are the most relevant sample annotation information files/tables for this type of analysis to be downloaded. In addition, does the GLASS consortium have the gene-level read count data for each sample or only the TPM values are available? From your experiences, what is the best normalization method between samples for TPM-based gene expression profiles? Thank you very much in advance. Best, Yiwen

Created by ywcus08
Hi Fred, Thank you very much for your information and great help. I wonder in your own analysis, whether you performed any normalization on the gene-level TPM values from different samples in the same batch. If so, what is the normalization method you used? Thanks. Yiwen
Hi Yiwen, Thank you for your interest. Please refer to the data dictionary when deciding which files/tables would be relevant to your analysis. It describes the purpose of each table in the Tables section, as well as their associated variables. All expression matrix data can be found in Files/current_release. We only have pseudocount data for transcripts, which was output by kallisto. kallisto outputs the pseudocounts and TPM for transcripts and we then converted that to TPM at the gene level. Please be advised that gene expression data was compiled from several different sources and as a result there is pretty heavy batch effect. We advise performing **paired comparisons** (such as paired t-tests) that compare each patient's initial tumor to their respective recurrent tumor. For a given patient, the initial and recurrent tumor will be in the same batch, so capturing these changes over time with paired tests should not require batch effect correction. Good luck with your analysis! Fred

Question about GLASS RNA-seq data page is loading…