Errors in RNA-Seq quantification affect genes of relevance to human disease

Table 2 Method performance on global simulated data

Method	Description	Pearson correlation coefficient
Star.htseq. u	STAR, HTSeq (union)	0.78
Star.htseq.ine	STAR, HTSeq (intersection-nonempty)	0.88
Star.htseq. is	STAR, HTSeq (intersection-strict)	0.86
Tophat.htseq. u	TopHat2, HTSeq (union)	0.78
Tophat.htseq.ine	TopHat2, HTSeq (intersection-nonempty)	0.87
Tophat.htseq. is	TopHat2, HTSeq (intersection-strict)	0.86
Star.cufflinks	STAR, Cufflinks	0.95
Star.cufflinks.mr	STAR, Cufflinks (multi-read-correct)	0.91
Tophat.cufflinks	TopHat2, Cufflinks	0.95
Tophat.cufflinks.mr	TopHat2, Cufflinks (multi-read-correct)	0.90
Sailfish	Sailfish (RPKM)	0.95
Sailfish	Sailfish bias-corrected (RPKM)	0.08

Pearson correlation coefficients between FPKM from each method and the expected FPKM from simulated data for 19,654 human protein-coding genes. RPKM reads per kilobase per million (see “Materials and Methods”)

ISSN: 1474-760X