7 months ago by
The tool version with problems is older. This older version is now outputting the extra bam file even if it is not specified to be output (this is a bug) -- the updated tool version processes this setting/output correctly.
Best tool version to use: htseq-count - Count aligned reads in a BAM file that overlap features in a GFF file (Galaxy Version 0.6.1galaxy3). The extra BAM output toggle is done with the parameter option "Additional BAM Output". If you choose to output this, a reference dataset is required -- either built-in "Locally cashed" or from the History (your own custom genome fasta).
If a BAM dataset has metadata problems (from any tool), it usually means that there was no output in most cases or there was an input format/content problem. This is part of the bug with the earlier tool version -- the BAM is output even when not specified, with no reference genome selection made, resulting in an empty result that cannot be indexed (leading to the metadata warning).
- Use the latest tool. This is the best solution. It includes bug fixes/enhancements and a restructured tool form.
- Use the older tool and either ignore or purge (permanently delete) the empty BAM output. It has no content, so does not count toward account quota usage.
- Use the older tool and set the option to output the BAM correctly (avoids the BAM metadata problem). The option is under "Set advanced options >> Set advanced options >> Additional BAM Output". Choose to output the BAM and the form will refresh allowing the selection of a reference genome.
Thanks, Jen, Galaxy team