Fig. 2

Inflation of tpm estimate in sub-sampled transcriptomes is consistently driven by variance in total number of reads mapped. A Consistent relationship between tpm estimate and total number of normalized reads is consistent with denominator decrease in shown equation for sub-sampled transcriptomes. This suggests that sub-annotated transcriptomes may suffer from similar inflation if samples are not corrected for total number of reads mapped. B Right skew tpm estimates of all examined selected genes reveals consistent inflation of the subsampled transcriptomes. Multi-modality of tpm values may be arising from competition for reads from paralogous targets only present in certain permutations of the sub-sampled transcriptomes (genes with distributions in red)