running MISO with ENSEMBL GTF if reads were aligned against UCSC genome

Question

[lavanya.kannan]

I aligned reads against UCSC genome using STAR, where UCSC gtf was used as reference annotation. As a result, I get the chromosomes in the format "chr1", .... When using pe-utils function in MISO, it is identifying 0 read pairs. I get this error when I tried using both Homo_sapiens.GRCh37.65.gff and hg19_ensGene.gff3 annotation files and also tried the --no-bam-filter option of the pe-utils function. I notice that my bam files also have "chrM". Will removing reads from "chrM" help?

Answer 1

[Devon Ryan]

Why not just fix the chromosome names?

1. samtools view -H foo.bam > header
2. Edit the header (mostly, remove "chr" and convert "chrM" -> "MT")
3. samtools reheader

[lavanya.kannan]

Devon, worked now. Thanks for the suggestion.

Answer 2

[Emily]

You can use the Ensembl File Chameleon to export an Ensembl annotation file with "chr" style names.

[Devon Ryan]

That's an interesting tool. Does it have the name conversions stored somewhere or does it just tack on chr everywhere?