DeepSNVMiner works on one data and dose not on the other

[vvsbiocode]

Hi, I would like to report a strange issue.
I tried DeepSNVMiner on human targeted DNA-seq data from Illumina NextSeq. The file samplename_R2.fastq.filter (inbut for aligning step) had some reads without quality string, as the result BWA reported:
[W::bseq_read] the 2nd file has fewer sequences.
[W::bseq_read] the 2nd file has fewer sequences.
[M::mem_pestat] skip orientation FF as there are not enough pairs
[M::mem_pestat] skip orientation FR as there are not enough pairs
[M::mem_pestat] skip orientation RF as there are not enough pairs
[M::mem_pestat] skip orientation RR as there are not enough pairs

and .sam file was empty.

When I did QC preprocessing on the same data and started DeepSNVMiner from block pool_reads, aligning showed the same error. Also, when I used cutadapt-preprocessed data on BWA separately from DeepSNVMiner, everything worked. Finally, when I did the same analysis with DeepSNVMiner on another data (from another project), aligning went well.

So, samplename_R2.fastq.filter file becomes damaged on the stage of pooling reads, I suspect. Nevertheless, I don`t know why on data from another project it worked well...

May I ask if there is any obvious solution that I can`t see?

The command that I used was, where variables are directories and numbers:
$DEAPSNVMINER_RUN -filename_stub $PREFIX -read1_fastq $OUTPUT_DIR/${samplename}${APENDIX1} -read2_fastq $OUTPUT_DIR/${samplename}${APENDIX2} -coord_bed $REGIONS_FILE -working_dir $OUTPUT_DIR -ref_fasta $INDEXED_GENOME_FILE -samtools $SAMTOOLS_LOCKAL -bwa $BWA_LOCKAL -uid_len1 $UMI_COMSEQ -no_adaptor -threads $THREADS -min_seqlen $MIN_LEN -sm_count $MIN_GROUP_UMI -sm_portion $MIN_GROUP_FRACTION -min_group $MIN_SUPERGROUP -start_command pool_reads

Thank you.