🧬 Join us in Boston for the 2026 bioBakery Workshop
(Cambridge, MA · July 20–24) —
Learn more & apply
🔬 We’re also delighted to welcome you to the
8th Annual HCMPH Symposium
(May 18, 2026) —
Translating the Microbiome: Turning Discovery into Implementation
View symposium details
I have followed the tutorials with test data and all worked well. I am currently running kneaddata on metagenome files, but end up with empty paired files as per below. I’ve updated bowtie2 to 2.5.4 and still get the same result. Has this something to do with fastq headers? mine start with:
12G Jun 6 12:26 02_R1_001_kneaddata.trimmed.1.fastq
238M Jun 6 12:26 02_R1_001_kneaddata.trimmed.single.1.fastq
11G Jun 6 12:26 02_R1_001_kneaddata.trimmed.2.fastq
134M Jun 6 12:26 02_R1_001_kneaddata.trimmed.single.2.fastq
0 Jun 6 13:05 02_R1_001_kneaddata_GRCh38_PhiX_bowtie2_paired_contam_1.fastq
0 Jun 6 13:05 02_R1_001_kneaddata_GRCh38_PhiX_bowtie2_paired_contam_2.fastq
0 Jun 6 13:05 02_R1_001_kneaddata_paired_1.fastq
0 Jun 6 13:05 02_R1_001_kneaddata_paired_2.fastq
12G Jun 6 13:39 02_R1_001_kneaddata_unmatched_2.fastq
12G Jun 6 13:39 02_R1_001_kneaddata_unmatched_1.fastq
6M Jun 6 13:39 02_R1_001_kneaddata_GRCh38_PhiX_bowtie2_unmatched_2_contam.fastq
4.2M Jun 6 13:39 02_R1_001_kneaddata_GRCh38_PhiX_bowtie2_unmatched_1_contam.fastq
12K Jun 6 13:40 02_R1_001_kneaddata.log
I updated my kneaddata to v0.12.4 and am still getting empty paired.fastq files when using paired end sequence data that I downloaded from SRA.
I changed my file headers to have the the following formats:
@ERR13326268.1.1
@ERR13326268.1#0/1
The original file headers seemed to be a problem, and the updates I made to the read headers did not fix my issue, and I am still getting empty paired files from kneaddata.
I downloaded kneaddata via mamba and can’t fix the python scripts used to build the program (as others have suggested).
Any help would be greatly appreciated! Thank you so much!
Hi. I don’t see any reason this wouldn’t work the way you’ve described it, even on an older version of kneaddata. The issue that was fixed was that certain formats added a space into the header that was not being parsed correctly. Are there any quality control issues? Do the headers match correctly across files? Are there repeated headers within the same file?
I figured out my problem which is that I accidentally changed the headers of the R fastq files (of PE samples) to /1, not /2. Once I did that, I fixed the issue and forgot to remove my comment above. Thanks for replying!