1
$\begingroup$

I'm trying to run RNA-seq on single-end mouse RNA data (50bp reads) but have very limited RAM (<4 gb) - I don't think I really have a way to work around this memory restriction, and wanted to try Kallisto's pseudoalignment (I conda installed it) as indexing and/or aligning with other tools like STAR and HISAT2 have resulted in the processes being killed. Building the index with kallisto index still results in the process being killed, so I downloaded the index files from https://github.com/pachterlab/kallisto-transcriptome-indices/releases/tag/v1 instead. I ran kallisto quant -i index.idx -o output --single -l 180 -s 20 sample1.fastq (I don't entirely know what the mean and sd arguments should be for my data), but this resulted in the following:

[quant] fragment length distribution is truncated gaussian with mean = 180, sd = 20
Error: incompatible indices. Found version 13, expected version 10
Rerun with index to regenerate

I've seen that other people have come across this error before when quantifying with kallisto, but it seems that they are building the index themselves while I was trying to use the pre-generated index from the unpacked mouse_index_standard.tar.xz file downloaded from the link above.

How can I get around this error? And in general, are there any suggestions on how to get raw/normalized count data from RNA-seq data with such memory limitations? I'm rather new to this.

$\endgroup$
1
  • $\begingroup$ If your finances at all allow, get more RAM. You're not soing much with such limited resources in the long run. $\endgroup$
    – ATpoint
    Commented Nov 30, 2023 at 22:49

1 Answer 1

3
$\begingroup$

It means that the version of the tool used for indexing and your current version are different and not compatible. Either downgrade your version to the version used for indexing, or build a new index with your current version. As by the GitHub repo, these indices that you link (supports kallisto versions >= 0.50.1), so use this version if you want to use the index.

If that does not work then I would open an issue at their GitHub.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.