Transcript quantification with RNA-Seq data

BMC Bioinformatics

Table 1 Correlation of underlying expression level and inferred abundances for different approaches

Approach	Correlation
	Across genes	Within genes (mean)
Position-wise inference with transcript profiles	0.820	0.635
Segment-wise inference with transcript profiles	0.693	0.488
Position-wise inference without transcript profiles	0.684	0.540
Segment-wise inference without transcript profiles	0.580	0.367

rQuant, which infers transcript abundances from read data at each position, is compared against a segment-based approach, which uses averages of averaged read counts at shared transcript segments. The correlation between true and inferred abundance was determined across all annotated transcripts; for alternatively annotated genes, the average of correlation within transcripts of each gene was calculated. We compare against not optimising the transcript profiles (i.e. uniform profiles).

ISSN: 1471-2105