Why don't the exon TPM values add up to ExonTPM?

[tspost]

Hi there,

Thank you for creating this very useful tool!

I don't understand why the sum of the individual exon TPM values in the genes.uni file don't add up to the ExonTPM value in the genes.out file for any given gene (or the sum of exon reads to ExonReads). Is that expected? The description of the model doesn't seem to explain that, but I may just be missing something.

Could you please explain (briefly)?

Thank you!

[r78v10a07]

Hi,
Thanks for using our tool.

Have a look at this figure:

TPM values in genes.uni are calculated only for features (exon or intron) that does not overlap with any other feature in the genome. This is the row at the bottom of the figure "Non overlapped features".

TPM values in genes.out are calculated without looking at any overlap of other genes. It is represented in the third row "add overlapping genes"

Therefore, it is correct that the sum you're mentioning does not agree as they count different genomic features.

Hope this help. Let me know if you need more info.
Roberto

[tspost]

Ah! Yes, that makes perfect sense. Thank you for explaining, especially so quickly!

[tspost]

Apologies, I have a follow-up question after all. Just to make sure I get this right: the TPMs for all exonic regions of a given gene are listed in the column "ExonTPM", i.e., the 10th column, of the genes.out file, correct? This would exclude any reads that fall into introns of the same gene but count every read that overlaps with any exon of the gene? The Output wiki lists this as the 14th column (https://github.com/ncbi/TPMCalculator/wiki/Output-files), so I just wanted to double-check.
Thank you!!