At least for the small data sets that we've tested, there is no difference in the count matrices using any of the --outSAMattrRGline or --readFilesManifest options.
But there is a difference in BAM files. Both the options produce the @RG header line. When --outSAMattrRGline is used, additionally, each alignment in the BAM file contains an RG tag, whereas --readFilesManifest does not add an RG tag. Shouldn't they produce the same result? Are RG tags used somewhere during the quantification?
to get RG tag into the BAM output you would need to add RG to --outSAMattributes, e.g.
--outSAMattributes NH HI AS nM RG.
With --outSAMattrRGline it is added automatically.
Hi @alexdobin, We're currently using
--readFilesManifest
to provide SmartSeq cell-ids information in the STARsolo Galaxy wrapper (https://github.com/galaxyproject/tools-iuc/blob/master/tools/rgrnastar/rg_rnaStarSolo.xml).At least for the small data sets that we've tested, there is no difference in the count matrices using any of the
--outSAMattrRGline
or--readFilesManifest
options.But there is a difference in BAM files. Both the options produce the
@RG
header line. When--outSAMattrRGline
is used, additionally, each alignment in the BAM file contains anRG
tag, whereas--readFilesManifest
does not add anRG
tag. Shouldn't they produce the same result? AreRG
tags used somewhere during the quantification?