Error in pbrun minimap2 with Parabricks 4.2.1-1 container

Healthcare & Life Sciences Parabricks
1

Hi,

I am using pbrun minimap2 with ONT reads with command:

docker run --gpus all --rm    \
       --volume $(pwd):/workdir --volume $(pwd):/output_dir \
        nvcr.io/nvidia/clara/clara-parabricks:4.2.1-1 \
        pbrun minimap2 \
        --preset map-ont \
        --index /workdir/workdir/Reference_genomic.mmi\
        --ref /workdir/workdir/Reference_genomic.fna \
        --out-bam /workdir/output_dir/ONT_minimap2_alignment.bam \
        --in-fq /workdir/all_ONT_long_Reads.fastq.gz

I get the following output:

[PB Info 2024-Jan-18 10:13:15] ------------------------------------------------------------------------------
[PB Info 2024-Jan-18 10:13:15] || Parabricks accelerated Genomics Pipeline ||
[PB Info 2024-Jan-18 10:13:15] || Version 4.2.1-1.beta4 ||
[PB Info 2024-Jan-18 10:13:15] || GPU-PBBWA mem, Sorting Phase-I ||
[PB Info 2024-Jan-18 10:13:15] ------------------------------------------------------------------------------
[PB Info 2024-Jan-18 10:13:15] Reading reference file.
[PB Info 2024-Jan-18 10:13:15] -------------------------------------
[PB Info 2024-Jan-18 10:13:15] Elapsed-Minutes Processed-Reads
[PB Info 2024-Jan-18 10:13:15] -------------------------------------
[PB Info 2024-Jan-18 10:13:15] 0.0 0
[PB Error 2024-Jan-18 10:13:16][src/reader.cpp:122] sizeOfSeq 114270 exceeds MAX_READ_SIZE 100000, expected sizeOfSeq < MAX_READ_SIZE, exiting.
For technical support visit Clara Parabricks v4.2.0 - NVIDIA Docs
Exiting…
Please visit NVIDIA Clara - NVIDIA Docs for detailed documentation

Is it possible to reset this value? I am guessing this error is popping up because it encountered a very long read in the read files.

Please advise.

2

I am having precisely the same issue with CLR (long) pacbio reads.

3

Given that minimap2 is supposed to handle long reads (that’s the whole reason for using it over bwa-mem), it isn’t great that it falls apart when a single long read comes along and exceeds 100kb. At the very least, it should be able to ignore enormous reads (1Mb+), but this parameter should be able to be set by the user as the technologies will keep increasing the read length over time.

4

We have found the issue and will have it fixed for the v4.3 release.

2 Likes
6

Works perfectly in ver 4.3.0-1.

Thanks for the fix!

1 Like