Importing mitochondrial RNA reads into SeqMonk

[millerd]

Hello community. I'm new to sequence analysis with SeqMonk and I must be doing something wrong because the software is ignoring mitochondrial RNA reads in SubRead RNA-seq output BAM files. I did find the instructions for placing an alias file in the same folder as the genome build but I still can't get it to work. I always get a 'couldn't extract valid data from chrM' error for each skipped entry. I'm assuming that SubRead uses the 'chrM' convention when mapping reads, so is it just the syntax or string in the alias file that I've got wrong? I can't look into the BAM files to check for myself.

Can someone tell what the correct string to use is?

Should it be chrM MT? (tab delimited) If so, that didn't work for me.

Sorry if this sounds a bit thick. I'm actually interested in the mitochondrial reads!

TIA David

[fkrueger]

I think your alias file should be the other way round:
MT chrM (tab delimited)

More info can be found here: http://proteo.me.uk/2011/06/adding-c...-into-seqmonk/

[millerd]

I think your alias file should be the other way round:
MT chrM (tab delimited)

More info can be found here: http://proteo.me.uk/2011/06/adding-c...-into-seqmonk/

Thanks for the tip; I'll try that. I did find the mapping page and followed the instructions, but wasn't clear which way round the convention should be. I'm also not absolutely sure what convention SubRead uses but assumed it is chrM based on SeqMonk's error reports. It's odd that SeqMonk should only stumble with the one chromosome when all others were imported correctly.

[fkrueger]

SeqMonk uses the Ensembl chromosome naming convention which is 1, 2, ..., X, MT while the UCSC one is chr1, chr2..., chrM. Seqmonk can cope with accepting the numeric chromosomes as 1 or chr1, but for MT there is actually a change from chrM to MT so it would involve a real change of the name. This is why it is failing. From what I heard Ensembl will soon adopt the UCSC nomenclature (including the chr) very soon, so then these mapping problems should sort themselves out naturally.