how to compare tophat output files with and without "_random" sequences

[EA01]

Hi,

I am trying to compare two RNAseq datasets obtained using Ilumina in two different dates.
I have the aligned files using Tophat but when I try to run Cufflinks I get the error message "Error: sort order of reads in BAMs must be the same".
I have checked the header of the aligned.bam files (with samtools view) and realized that one of the datasets had sequences aligned to a genome containing sequences which could not be specifically located on a chromosome (these are sequences called chr6_random, chr9_random etc. etc.) and the other did not. This means that one file has some sequences aligned to a location which is not present at all in the other and this is why Cufflinks is not able to work.

Does anyone know what is the problem with my files and have any suggestion on how to fix this without remapping?
I have actually tried to remap some of the files but I am still getting the same type of aligned sequences in each case...

Thanks!
EA

[bruce01]

You can just remove the lines with the different chromosome_ID using samtools:

samtools view aligned.bam chr1 chr2 ... chrN > aligned.chr1-N.bam

BAM needs to be coordinate sorted (samtools sort), indexed (samtools index).

Did you use the same reference to align to? You could remove these 'random' chromosomes/contigs from the reference so you just have the 'consensus' chromosomes (this presumes you are working on a well enough annotated species).

[EA01]

Thanks! I'll try this.
Yes, I'm using the mm9, so it should be well annotated.