Invalid BAM file as an input to rgi kmer_query

[monalibhanja]

Input features:
Paired metagenomics reads, preprocessing done by cutadapt and fastp.
RGI features: RGI v6.0.5; CARD database v4.0.1
Command used for rgi bwt-
rgi bwt --read_one sample_R1.fastq.gz --read_two sample_R2.fastq.gz --output_file sample_rgi_bwt --include_other_models

This was done using default rgi bwt options, hence KMA was used as the aligner.

Output file "sample_rgi_bwt.sorted.length_100.bam" was used as an input for the next step. Command for rgi kmer_query-
rgi kmer_query --bwt --kmer_size 61 --minimum 10 --input sample_rgi_bwt.sorted.length_100.bam --output sample_rgi_kmer

Output:
ERROR 2025-12-09 15:01:57,879 : failed to parse BAM file: /home/car-nmgd/main_samples/HWW_samples_24/rgi_kmers/rgi_outputs_C03/C03_S1.sorted.length_100.bam, please check which aligner software was used to produce the BAM file in the RGI BWT step.

What BAM file should be used as the input for rgi k-mer command? Should I be using another aligner to generate the BAM file?

[mawlodarski]

Hi Monali, can you attach the two files (R1 & R2) for testing? Thank you.

[raphenya]

@monalibhanja @mawlodarski rgi bwt both take in compressed and uncompressed fastqs. Maybe the files are not compressed properly?

[raphenya]

@monalibhanja Ok I did the following tests and no errors:

# local db testing for rgi 6.0.5
# ==============================

# load data
rgi auto_load --debug --clean --local > load.log 2>&1 &

# run rgi bwt
rgi bwt --read_one 10_R1.fastq.gz --read_two 10_R1.fastq.gz --output_file sample_rgi_bwt --include_other_models --local --debug --clean > 1.log 2>&1

# run rgi kmer_query
rgi kmer_query --bwt --kmer_size 61 --minimum 10 --input sample_rgi_bwt.sorted.length_100.bam --output sample_rgi_kmer --local --debug > 2.log 2>&1


# system wide db testing for rgi 6.0.5
# ==============================

# load data
rgi auto_load --debug --clean > load-systemwide.log 2>&1 &

# run rgi bwt
rgi bwt --read_one 10_R1.fastq.gz --read_two 10_R1.fastq.gz --output_file sample_rgi_bwt --include_other_models --debug --clean > 3.log 2>&1

# run rgi kmer_query
rgi kmer_query --bwt --kmer_size 61 --minimum 10 --input sample_rgi_bwt.sorted.length_100.bam --output sample_rgi_kmer --debug > 4.log 2>&1

[raphenya]

@monalibhanja @mawlodarski Ok, there is an error parsing due to header changes. I will try to fix this ASAP. Cheers.

[monalibhanja]

@raphenya I tried running rgi bwt with both compressed and uncompressed, but I came across the same "please check which aligner software was used to produce the BAM file in the RGI BWT step" error.

I also tried using the same commands as you mentioned in "system wide db testing for rgi 6.0.5", again the same error.

[raphenya]

@monalibhanja Yes, I'm working on a solution. Cheers.

[raphenya]

@monalibhanja Which version of KMA are you running?

[monalibhanja]

@raphenya I am using version "KMA-1.6.7".

[raphenya]

@monalibhanja So, some versions of KMA don't set @PG flag. I have tested as follows:

KMA version	Yes - works; No - doesn't work
1.6.10	Yes
1.6.9	Yes
1.6.8	Yes
1.6.7	No
1.6.6	Yes
1.3.28	No

Ref: https://bitbucket.org/genomicepidemiology/kma/issues/54/feature-request-write-pg-flag-in-sam

Use the following command to check for the flag after running RGI-bwt

samtools view -H output.sorted.length_100.bam | grep "@PG"

Results:

@PG	ID:KMA	PN:kma	CL:kma -mem_mode -ex_mode -1t1 -vcf -ipe ...
...

Hope this helps. Cheers.