Question about running Cufflinks on ENCODE RNA-Seq data

[juliacc]

Hello,

I'm new in the community but I found many helpful posts in the past, so here is my question.

I downloaded around 60 .bam files from the ENCODE repository. The header size was giving problems and I converted them to .sam as one of the answers in these thread suggested.

samtools view -h -o file.sam file.bam

And when I run Cufflinks I get many lines with this error:

SAM error on line 862765: found spliced alignment without XS attribute

I am executing this on batch, so it is still running and, for some files, Cufflinks finish and produces the gene count, for other files I am not sure if it just hasn't finished yet.

I've been reading many posts about the XS attribute, but I don't know why would it be missing if I just converted from .bam to .sam. And would you know if it affects the read counts for genes?

Thank you,
Julia

[AmitPande]

Did you find a solution yet ? Even I am facing the same prolem with ENCODE RNAseq data

[bye]

Not sure how this error will affect the counts, but it's not hard to add XS tags.

samtools view -Xh yourfile.bam | awk -F"\t" '{if(NF<7)print $0; if($2~"P"&&$2!~"r"){print $0"\tXS:A:+"}else if($2~"P"&&$2~"r"){print $0"\tXS:A:-"}else{print $0}}' > yourfile.xs.sam

XS tags label the strand of alignment; samtools view -X will convert the flags into human readable format, then you can add strand information accordingly.