Fig. 1

RNA-seq read coverage score shows between-sample heterogeneity in the pattern along transcripts. a The read coverage score, defined as the number of reads that cover each base pair, is expected to have a trapezoidal shape along the transcript if the read start position is uniformly distributed. b An example from SEQC-AB data shows a non-uniform but consistent read coverage pattern, where the average magnitude of coverage score for each sample may faithfully represent the transcript abundance given the sequencing depth is normalized. The diagram in red under the coverage plot shows the total transcript with exon boundaries from genome annotations (same for c, d and e below). c The TMEM229B gene from the breast tumor data shows differential coverage score patterns between FF and FFPE samples. Reads from FF samples are continuously distributed across the entire transcript while those from FFPE samples are highly enriched in a few disjoint blocks or fragments. d The ACTN4 gene from the GBM data with RIN number = 10 vs. RIN = 4 shows clear degraded coverage score towards the 5′ end of the transcripts in the latter group. e An example from the SEQC-AB data shows that alternative splicing likely causes sharply decayed coverage score across the entire alternatively spliced exon region