BiocParallel error when using Bambu in quantification mode

[apsteinberg]

Hi Andre and the Bambu team,

I am encountering an error when I am running Bambu in quantification mode on some of my samples. After our discussions in issue #448, I decided to try to include all detected transcripts in my analysis as NDR < 0.4 for all of the transcripts. I am finding however that for some of my samples now I am encountering a BiocParallel which reads:

Error: BiocParallel errors
  1 remote errors, element index: 1
  0 unevaluated and other errors
  first remote error:
Error in if (annotatedIntronNumberNew > annotatedIntronNumber & !is.na(annotatedIntronNumber)) {: missing value where TRUE/FALSE needed
In addition: There were 15 warnings (use warnings() to see them)

The commands I am using to run Bambu are:

annotations <- prepareAnnotations(gtf.file)

create_directory(outdir)

se.quantOnly <- bambu(reads = input_bam,
                      annotations = gtf.file,
                      genome = fa.file,
                      discovery = FALSE,
                      verbose=TRUE,
                      lowMemory=TRUE,
                      yieldSize=1e6)

writeBambuOutput(se.quantOnly, path = outdir)

se_out <- sprintf("%s/se.quantOnly.RData",outdir)
save(se.quantOnly, file = se_out)

I am attaching the full log here in case it is helpful:

quant_biocparallel_error.zip

This was run with bambu 3.5.1. Any clues as to what's going on here?

Thanks, Asher

[cying111]

Hi @apsteinberg ,

Sorry for the late reply!

Based on the error message, it looks like a few samples had zero reads aligned to the reference genome. This could be due to data quality issues or a mismatch between the genome FASTA and the reads (e.g., seqname style mismatch or out-of-range sequences). You'll need to investigate further to pinpoint the cause.

I recommend pre-checking each sample and using only those with at least some reads.

Hope this helps! Let me know if you have any other questions. Thank you Warm regards, Ying

[hmutpw]

Hi @apsteinberg ,

Sorry for the late reply!

Based on the error message, it looks like a few samples had zero reads aligned to the reference genome. This could be due to data quality issues or a mismatch between the genome FASTA and the reads (e.g., seqname style mismatch or out-of-range sequences). You'll need to investigate further to pinpoint the cause.

I recommend pre-checking each sample and using only those with at least some reads.

Hope this helps! Let me know if you have any other questions. Thank you Warm regards, Ying

Hi @cying111 ,

I also encountered the same error when the "lowMemory=TRUE" is setted. However, when I set "lowMemory=FALSE", the error information disppared. I am pretty my 99% percent of the reads were aligned to ref genome by minimap2, so, I guess that if some chromosomes do not have any reads mapped, the error may occour because lowMemory process the data by chromosome? Also I saw the same situation at https://github.com/GoekeLab/bambu/issues/439. Maybe this can be fixed in the next version.