kallisto pseudoaligned bam files to IGV

Question

[hothriananya]

Hi, has anyone found the solution for this? I am running kallisto with --genomebam option along with the --chromosomes, The bam file generated is already sorted by kallisto and indexed so I think I do not have to sort it again. I uploaded these bam files to the IGV and i cant see any alignment and it shows there are 0 counts. I did downstream analysis with abundances.h5 files and differential expression looks fine. I am posting my commands below

"${KALLISTO} quant -i ${GENOME}/hg19.idx -o "MB-7249-Sample-1" -t 10 --genomebam --gtf ${GENOME}/hg19.gtf --chromosomes ${GENOME}/hg19_chromsizes.txt""

Any suggestions on why pseudo-aligned bam files cant be viewed in IGV

Thank you, Hothri

[hothriananya]

Hi, Thank you for your suggestions, I figured out why I cant view pseudobam files in IGV. I have to upload cdna fasta file to the igv to view bam files. I also tried --genomebam option but it didnot work and i am figuring out why.

The only drawback of viewing transcript pseudobam is I can only view one transcript at a time rather than whole gene. which is bit annoying but I guess this issue will be solved with the bam files are genrated with --genomebam option. I am trying to fix this and will get back when I figure it out. Thank you all for suggestions.

[Kristin Muench]

Hello! Did you have any luck with this issue? I'm having the same problem.

[hothriananya]

Hi, Yes, I identified its an alignment problem. I re-built Kallisto index with cDNA fasta file and realigned the fastqs with genome option and now I can view the whole genome.

[Sergio]

Hi, can you share how did you realigned the fastq? I'm also using a cDNA index for kallisto quant with the --genomebam, --gtf and --chromosome options but my BAM files don't contain chromosome information. Therefore they can't be viewed using the whole genome but only if I upload the cDNA to IGV (which as you said is annoying because you can only see one transcript at a time).

Answer 1

[pmelsted]

There could be a mismatch between chromosome names.

Can you show the first 10 lines of hg19.gtf and hg19_chromsizes.txt as well as the names of the chromosomes you see in IGV (these are taken from the FASTA file provided for the genome). If the names do not match IGV will not display anything.

My guess is that the GTF file has chr1 whereas IGV is using 1, etc. The solution is to fix one of those names.

Answer 2

[pbpanigrahi]

Since we don't have access to your data, here are a possible check sums

I am assuming hg19.idx is the index generated using transcriptome sequences and not the reference genome since kallisto aligns reads against transcriptome (not genome).

There are 2 possible bam options, --pseudobam or --genomebam. Former is bam against transcriptome while later against genome.

If former you want to visualize in igv, then make sure bam index file (bai) is available in the same directory as bam. e.g. pseudoalignments.bam and pseudoalignments.bam.bai

If later you want, then u need gtf and chromosome length file, which in the command you have already given. So please check whether chromosome names in gtf file (-gtf) and chromosome file (-chromosome) matches.

Manual says "For a larger transcriptome we recommend downloading the GTF file from the same release and data source as the FASTA file used to construct the index." so check whether the chromosome information in hg19.idx files and gtf matches.

Hope you find solution