outSAMstrandField intronMotif

[Felix Schlesinger]

Hi,

I ran into an issue with outSAMstrandField intronMotif. It seems to filter out some spliced alignments that look fine to me.

For example for these reads, aligned to hg19 (with refseq annotation):

@foobar

AGGTCTCTGTTCAGGTCAACGTCTCCTTCCCGGCCAGTGTGCAGCTGCACACGGCGGTGGAGATGCACCACTGGT

+

CBBCCGG>GGGGGDGGGGGGGGGGGGGGG1FGGGGGFGGGGGG<GGGGGG1GEGGGDFGGFBGGGFGGGGGGGDG

@foobar

GTTGACGTGGGTGGGCTGGTTGAGGCGCAGACACCCGTGCCGCACGGTCTCATTGGCTGCCGGCTCCAGGAACTC

+

CBCBCGGG0GGGGGGGGGGFFGGGGGGGGGGGGGGGGGGGGGGCG1GGGGGGFGGGGGBGGGGFFGGGGGGGFGG

without outSAMstrandField intronMotif STAR finds a spliced alignment:

foobar  99      chr1    156841521       255     27M1877N48M

with outSAMstrandField intronMotif the read is instead softclipped at the exon boundary.

foobar  99      chr1    156843425       255     27S48M

The junction is canonical GT/AG on the + strand.

Any ideas? Am I missing something?

Felix

[Alexander Dobin]

Hi Felix,

I have tried these reads with hg19/Gencode19, and it did not have the problem you described:

@CO     user command line: /sonas-hs/gingeras/nlsas_norepl/user/dobin/STAR/Releases/STAR_2.4.2a/bin/Linux_x86_64/STAR --genomeDir /sonas-hs/gingeras/nlsas_norepl/user/dobin/STAR/SandBox/ENCODEpipelines/STAR_RSEM_tests/Prep/STARgenomeMale/ --genomeLoad LoadAndKeep --readFilesIn r1 r2 --outSAMattributes jM jI --outSAMstrandField intronMotif

foobar  99      chr1    156841521       255     27M1877N48M     =       156843565       2119    AGGTCTCTGTTCAGGTCAACGTCTCCTTCCCGGCCAGTGTGCAGCTGCACACGGCGGTGGAGATGCACCACTGGT     CBBCCGG>GGGGGDGGGGGGGGGGGGGGG1FGGGGGFGGGGGG<GGGGGG1GEGGGDFGGFBGGGFGGGGGGGDG     jM:B:c,21       jI:B:i,156841548,156843424      XS:A:+

foobar  147     chr1    156843565       255     75M     =       156841521       -2119   GAGTTCCTGGAGCCGGCAGCCAATGAGACCGTGCGGCACGGGTGTCTGCGCCTCAACCAGCCCACCCACGTCAAC     GGFGGGGGGGFFGGGGBGGGGGFGGGGGG1GCGGGGGGGGGGGGGGGGGGGGGGFFGGGGGGGGGG0GGGCBCBC     jM:B:c,-1       jI:B:i,-1       XS:A:+

This junction is annotated in Gencode19.

Could you please send me your Log.out file, and also the link to the GTF?

Cheers

Alex

[Felix Schlesinger]

I cannot reproduce this when I regenerate the genome index with the latest STAR release (and the same annotation). Not sure what the root cause was, but looks like it got fixed in the meantime anyway.

Thanks