fastq-multx not identifying barcodes

[SJRussell]

I have a fastq file with 1.7 million reads of 75 bps, and 28 different barcodes. My command:
``fastq-multx -l barcodes.txt all_samples.fastq -e -o %.fq -x"

Part of my barcode file: barcodes.txt:

Head of my fastq, with some barcodes highlighted:

My output, which leaves most reads multiplexed:

Suggestions? What does (shifted) mean?

Thanks,

Stewart

[benligan]

Having this problem as well, for some reason fastq-multx is not recognizing the barcode correctly.

[drsuuzzz]

Can you give some more information about the type of assay and sequencing equipment used?

[benligan]

Sure, Illumina Miseq, paired ends. R1 forward and R3 reverse sequences.

/Users/###/miniconda2/bin/fastq-multx -B /Volumes/homes/###/SRA/Columbia.Gut.Murine/map/reversecomplement/reversecomplement.txt /Volumes/homes/###/SRA/Columbia.Gut.Murine/2_fastq/lane1_NoIndex_L001_R1_001.fastq /Volumes/homes/###/SRA/Columbia.Gut.Murine/2_fastq/lane1_NoIndex_L001_R3_001.fastq -o /Volumes/homes/###/SRA/Columbia.Gut.Murine/SRA/R3/R3.%.fastq /Volumes/homes/###/SRA/Columbia.Gut.Murine/SRA/R3/R1.%.fastq

/Users/###/miniconda2/bin/fastq-multx -B /Volumes/homes/###/SRA/Columbia.Gut.Murine/map/forward/forward2.txt /Volumes/homes/###/SRA/Columbia.Gut.Murine/2_fastq/lane1_NoIndex_L001_R1_001.fastq /Volumes/homes/###/SRA/Columbia.Gut.Murine/2_fastq/lane1_NoIndex_L001_R3_001.fastq -o /Volumes/homes/###/SRA/Columbia.Gut.Murine/SRA/R1/R1.%.fastq -o /Volumes/homes/###/SRA/Columbia.Gut.Murine/SRA/R1/R3.%.fastq
Using Barcode File: /Volumes/homes/###/SRA/Columbia.Gut.Murine/map/forward/forward2.txt

Returns:

1. Empty fastq files

I know each of these samples should have 5000-10000 amplicons per sample.

The behavior is also very erratic, it works with some fastq files (demultiplexes appropriately) and on other runs it doesn't demultiplex at all and returns empty files.

Primer barcodes.

[benligan]

I am also using: Version: 1.3.1

Here is cat of the fastq file

+
AAAAAFFAA11>AEGGGFGCGGHGGGAEHHFHHGHHEGGGHH1B01BF/BFEGCEE@/B2222BFFFF1BEFACEHF1B@/EEGEE<FHH1F00?0<B11<//?>///1?/C////1?F0<0>FCGFF=<GEGG<-..<<E00:CGHHHH:
@MISEQ01:40:000000000-ART6C:1:1101:16675:2021 1:N:0:
GAAATATCCTTTGCAGTAGCGCCAATATGAGAAGAGCCATACCGCTGATTCTGCGTTTGCTGATGAACTAAGTCAACCTCAGCACTAACCTTGCGAGTCATTTCTTTGATTTGGTCATTGGTAAAATACTGACCAGCCGTTTGAGCTTGAG
+
AAA3AFFFFFFDGG54BDEGGCGGFFHHGGFCFHFHGHHDHHHGGGGDHHHHHHGGGGGHHFFGHHHHHHHGHHHFHHHHHHHFGGHGHHHHHHGGGGHGFHGHHHHGGHHHHHFHHGHGGHHGHHHHHHGHHFHHHGGGGHFHGFHHHEG
@MISEQ01:40:000000000-ART6C:1:1101:19362:2021 1:N:0:
TACGTAGGGGGCAAGCGTTATCCGGATTTACTGGGTGTAAAGGGGGCGCAGACGGCAATGCAAGCCAGGAGTGAAAGCCCGGGGCCCAACCCCGGGACTGCTCTTGGAACTGCATGGCTGGAGTACAGGCGGGGCAGGCGGAATTCCTAAT
+

[ExpressionAnalysis]

Would you be able to attach some sample data that we can work with to recreate the problem?

[J-Sabino]

I am having a similar problem.
I have the same samples in 2 lanes (Hiseq Illumina), in total 992 samples. In lane 1 it returns 983 samples and in lane 2 it returns 985 samples. This is strange because I checked and the barcode sequences are there, but fastq-multx is not returning everything as needed.

[wltrimbl]

João, can you prepare subsets of the data files (perhaps 20 lines of all of the relevant fastqs?) and the sample mapping file that will produce the bad output? (Short, self-contained example of the bug)

[J-Sabino]

Hi, I was having a problem in generating this data subset and I am not allowed to share the full dataset.

In the meanwhile, I used QIIME2 to demultiplex the samples and it worked fine. I realized that the low read samples were the ones not being picked with fastq-multx. I need to say that I always used fastq-multx in Miseq and it worked fine. So maybe it is something with Hiseq or the size of the dataset.

Again, I am sorry for not being able to provide a self-contained example of the bug.