Can't find left Boundary

[cliftonlewis]

Is the bug primarily related to salmon (bulk mode) or alevin (single-cell mode)?
alevin
Describe the bug
A clear and concise description of what the bug is.
I am running the following command on some sci-rna-seq3 samples and it seems to not work as expected.
salmon alevin -i af_splici/dm6_splici_idx/ -l ISR -1 data/SRR17122012_1.fastq -2 data/SRR17122012_2.fastq -o SRR17122012 --tgMap transcriptome_splici_fl52/transcriptome_splici_fl52_t2g.tsv -p 28 --sciseq3 --justAlign
I then took the output into alevin-fry to create a generate-permit-list and it gives me the error that salmon hasn't added the extra bps to account for the chemistry
"thread 'main' panicked at 'assertion failed: (left == right)
left: 20,
right: 19: found barcodes of different lenghts 20 and 19', src/cellfilter.rs:203:13
note: run with RUST_BACKTRACE=1 environment variable to display a backtrace"
Thus I re-ran salmon alevin without the --justAlign flag and it seems to hit a different error
"### alevin (dscRNA-seq quantification) v1.9.0

[ program ] => salmon

[ command ] => alevin

[ index ] => { af_splici/dm6_splici_idx/ }

[ libType ] => { ISR }

[ mates1 ] => { data/SRR17122012_1.fastq }

[ mates2 ] => { data/SRR17122012_2.fastq }

[ output ] => { SRR17122012 }

[ tgMap ] => { transcriptome_splici_fl52/transcriptome_splici_fl52_t2g.tsv }

[ threads ] => { 28 }

[ sciseq3 ] => { }

[2022-11-28 21:13:57.772] [alevinLog] [info] Found all transcripts to gene mappings
[2022-11-28 21:13:57.781] [alevinLog] [info] Processing barcodes files (if Present)

processed 10 Million barcodes

[2022-11-28 21:14:01.454] [alevinLog] [info] Done barcode density calculation.
[2022-11-28 21:14:01.454] [alevinLog] [info] # Barcodes Used: 1 / 10285890.
[2022-11-28 21:14:01.455] [alevinLog] [error] Can't find left Boundary.
Please Report this issue on github."

Specifically, please provide at least the following information:

• Which version of salmon was used? salmon (1.9.0); alevin-fry (0.8.0)
• How was salmon installed (compiled, downloaded executable, through bioconda)? Conda install
• Which reference (e.g. transcriptome) was used? Drosophila melanogaster (BDGP6.32 (GCA_000001215.4))

Desktop (please complete the following information):

• OS: Ubuntu Linux

Thanks in advance for all your help regarding this!

[rob-p]

Pinging @Gaura here — any idea what might be causing the inability to properly process the barcodes?

@cliftonlewis — would you be able to share a sampling of the reads for us to examine and help debug with?

[Gaura]

@cliftonlewis: could you tell us version of alevin-fry are you using?
@rob-p: The cellbarcode length should be 21. It is variable b/w 19 or 20 so AC or A is added to make it 21. It could be the odd-even error we saw on previous version of alevin-fry. Wrt the run without --justAlign, I would need to take a closer look.

[cliftonlewis]

Hi,

Thanks for your rapid reply. So I am using version 0.8.0. All the fastq files were deposited under SRP349178

Here is just a snippet, is this informative? What else can I supply to help you guys.

@SRR17123283.1 1 length=34
TAGAANCAATCAGAGCGGGTTGGTCATCTCNNCN
+SRR17123283.1 1 length=34
AAA/A#EEEEEAE<EEEEEEEEEEEEEEEE##E#
@SRR17123283.2 2 length=34
AGGCGNAATCAGAGATCCTAGGGCCGCCGCNNAN
+SRR17123283.2 2 length=34
AAAAA#EEEEEEEAEEEEEEEEEEEEE//A##/#
@SRR17123283.3 3 length=34
GTCGTNACTCAGAGAATATTCTTTCTAATAGNAN
+SRR17123283.3 3 length=34
/6AA/#EAEEAEEAEEEEE/AEEEAEAEEEE#E#

[rob-p]

@Gaura: So, it seems the problem is during barcode extraction and mapping — we're not even getting to the step where alevin-fry is being run, right? Does anything look suspicious in the read snippet above (or in the corresponding read set that @cliftonlewis mentioned)?

[cliftonlewis]

Hi, just wondering if you guys managed to understand what the source of my error was? I don't know if I just don't know how to work the sci-rna-seq3 data

[Gaura]

Hi @cliftonlewis,

Sorry for the delay. I tested it on another file and it worked fine. I would like to look at some info from your file. Could you:

1. Post the salmon log of the first run, the one that you did with --justAlign
2. There should be a map.rad file in your output directory (SRR17122012). Can you run the command: alevin-fry view --rad map.rad > rad.txt and share the rad.txt file?

Thanks.

[Gaura]

Hi @cliftonlewis and @rob-p,

I figured out the issue. It was with the length check in the extract barcode code. I can push a fix today.

[Gaura]

Thanks for reporting it @cliftonlewis. I have tested the fix and it works both with and without rad mode (--justAlign). The solution is in pr #817.

[cliftonlewis]

That's cool. Do you know when the change would become a commit?

[rob-p]

Sorry I missed this! It's been on that branch (and committed) since Gaurav's PR. It's now been merged into master and included in the latest release (1.10.0).