Post-error behaviour: bamtofastqerror: Didn't find both records for a paired end read. Is your BAM file complete?

[endeneon]

Hi I am using the bamtofastq to dump cellranger-arc generated atac-seq bam files (and barcode-subsetted by 10x subset-bam) to their original fastq format (_I1_001, _R1_001, _R2_001. _R3_001) since I need to re-align the R1_001 and _R3_001 as Read1 and Read2 using bowtie2 (which I have successfully used before). However, during the dumping process, the bamtofastq threw out an error message as:

bamtofastqerror: Didn't find both records for a paired end read. Is your BAM file complete?
However, when I go to the output directory and check the outputs, It seems that there were several sets of fastq files. So my question is: what is the behaviour of bam2fastq after throwing out that error message? Does it 1) terminate itself immediately or 2) ignore that unpaired read and proceed to the next? If (2), then the fastqs are still usable but (1) would surely produce truncated fastq files. What case would it be? Thanks!

[blake-bowen]

Did you figure out a solution for this? I'm having the same issue.