How to perform corrected Cell Barcode/UMI calling with an externally sorted BAM?

Question

[mk]

I require Cell Barcode / UMI calling but am unable to sort BAMs within STARsolo due to a limited system RAM. When I set limitBAMsortRAM==<ammount of RAM on my system>, STAR throws the following error:

EXITING because of fatal ERROR: not enough memory for BAM sorting: 
SOLUTION: re-run STAR with at least --limitBAMsortRAM 273246110398

Proposed solution:

1. Run STARsolo with
2. run samtools sort with my hard memory limits as follows:

   samtools sort -m ${BAMSORTRAM} -@ ${BAMSORTNTHREADS} -o "${base}_STARsolo/${base}_Aligned.sortedByCoord.out.bam" "${base}_STARsolo/${base}_Aligned.out.bam"
   

So how do I re-run STARsolo just to call Barcodes/UMIS on:

1. the original alignment (which produced the unsorted BAM subsequently aligned with samtools above) and the subsequently sorted BAM
2. the sorted BAM from samtools (command shown above)

[GenoMax]

If you can't make starsolo work with hardware you have then perhaps look into alevin-fry instead: https://combine-lab.github.io/alevin-fry-tutorials/2023/simpleaf-piscem/

[dsull]

Try setting a high number for outBAMsortingBinsN.

Also, use the latest version of STAR.

Finally, your FASTQ files might have already been pre-sorted, so you can try shuffling the reads in your FASTQ files.

STARsolo shouldn’t take almost 300 gb of memory to sort a BAM file… I’m fairly certain that doing some or all of the above will fix it.