HISAT2 error failed to read header from "-"

Hello today i was using HiSAT2 to fo a mapping of my fastq paired-end secuences to the reference genome of my organism (Hydractinia symbiolongicarpus) and the software returned this error:

Total time for call to driver() for forward index: 00:07:18
(ERR): mkfifo(/tmp/39861.inpipe1) failed.
Exiting now …
samtools sort: failed to read header from “-”
[main_samview] fail to read the header from “-”.

I dont know exactly what it means. This is the link to the history im working on:

Hi @Nicolas_Romero_Villa
I successfully tested HiSAT2 with a custom genome. I don’t know why your jobs have failed. I converted your genome and reads into plain text formats, fasta and fastqsanger, and completed a test job with 1k PE reads. It is possible that the reads were compressed in non-standard option, and HiSAT2 cannot unpack the data. However, I tested only 1k reads, so there might be an issue with the data. It is very unusual situation.
Just in case, procedure for conversion of formats: click at Edit Attributes (pencil icon) > in the middle window switch to Convert tab and select an appropriate option from the pull-down menu.
Kind regards,

Hi @igor @Nicolas_Romero_Villa
I ma having the same problem but with the human hg38 genome. The same data works fine on galaxy.eu so I assume it’s a problem with this server. I have reported it as a bug.

Hi @igor i converted the data as you suggest and i got the BAM file from HISAT2 the only cuestion that i have now its if the free space that Galaxy gives its enough to all my work flow
Thank you very much

Ho @Nicolas_Romero_Villa
I am glad it works for you. Looks like a compression issue. It is very unusual. You can try one sample in Europe, because of information shared by @ks-bris. Files can be copied to another Galaxy server directly, using link (chain icon). Paste the links into Galaxy Upload menu, Past/Fech data tab.
Map reads in batches and delete uncompressed files after completion of mapping jobs. BAM files contain both mapped and unmapped reads. You’ve done QC of raw data (FastQC), so you probably don’t need reads anymore in Galaxy.
I am not admin on ORG, and I don’t know the current quota policy. On Galaxy Australia we recommend analysis in batches.
Kind regards,

1 Like