DESeq: Read counts vs. BP counts

burkard · 08-06-2010, 12:52 AM

Hi Simon,

is it possible to use raw bp counts (how often a bp is sequenced within a region) of an exon/transcript for differential expression analysis with DESeq?
First I think, that if we consider the expression of an individual exon the bp count would avoid double counting of split reads (especially if it belongs to a major part to one exon). It would only assign a portion of the read to one exon. Second, it would be nice to use with cufflinks where cov*length = bp count. Also for cufflinks the bp count divided by the read length is not necessarily equal the read count, especially for alternative transcripts or if you provide a gtf of known genes (even though its an approximation).

Thanks a lot and best regards,
Thomas

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
read counts per transcript, edgeR, tophat	chrisbala	Bioinformatics	46	02-27-2015 05:51 AM
Converting FPKM from Cufflinks to raw counts for DESeq	jebe	Bioinformatics	34	02-05-2014 09:19 AM
Raw read counts for RNAseq	biofreak	Introductions	13	01-16-2013 06:28 AM
Raw read counts for RNAseq	biofreak	RNA Sequencing	2	06-15-2011 06:56 AM
miRNA read counts regarding DE	vebaev	RNA Sequencing	16	05-19-2011 12:52 PM

08-06-2010, 12:52 AM	#1
burkard Junior Member Location: Vienna Join Date: Jul 2010 Posts: 2	DESeq: Read counts vs. BP counts Hi Simon, is it possible to use raw bp counts (how often a bp is sequenced within a region) of an exon/transcript for differential expression analysis with DESeq? First I think, that if we consider the expression of an individual exon the bp count would avoid double counting of split reads (especially if it belongs to a major part to one exon). It would only assign a portion of the read to one exon. Second, it would be nice to use with cufflinks where cov*length = bp count. Also for cufflinks the bp count divided by the read length is not necessarily equal the read count, especially for alternative transcripts or if you provide a gtf of known genes (even though its an approximation). Thanks a lot and best regards, Thomas