Hello, I'm trying to run featureCounts (v2.0.6) on mapped BAM files from metagenomic data with the following code:

featureCounts -p -F SAF -T 20 -a 8.prodigal/dereplicated_contigs.SAF -o 8.featureCounts/contigs_featureCounts.txt 8.mapping/*.sorted.bam

The BAM files are roughly 1.5GB and the SAF file is 2.5GB. featureCounts loads in the annotation SAF file with no issues but as soon as it comes to processing the BAM file I get the following error:

|| Process BAM file S1.sorted.bam...                                          ||
featureCounts: hashtable.c:309: HashTableCreate: Assertion `numOfBuckets > 0' failed.
/var/spool/slurm/job48387597/slurm_script: line 17: 155999 Aborted     featureCounts -p -F SAF -T 20 -a 8.prodigal/dereplicated_contigs.SAF -o 8.featureCounts/contigs_featureCounts.txt 8.mapping/*.sorted.bam

I've checked to see if the BAM file was corrupt but everything seemed fine. I tried adjusting CPU and memory but that hasn't worked either. Any suggestions/advice would be great! Thanks

Hi Jessica, thanks for sharing the data with us! We have identified the issue behind the error: the header is very long in the BAM file. In fact, the length of the header is very close to the upper limit in BAM specification (2^31 bytes). Because featureCounts couldn't correctly allocate memory for storing the header, an error was reported.

We have fixed this issue and made a new version (https://sourceforge.net/projects/subread/files/subread-2.0.7/). The new version can correctly process the BAM file and the SAF file that you shared with us.