BETA Work in progress — fidelity to the original filter is not guaranteed. Check your results.
web-fastq-human-remover Remove human (or any reference) reads — entirely in your browser

Remove human reads from sequencing data

Host DNA contaminates most metagenomes. Archives require it gone before submission; ethics and GDPR rules often require it gone before sharing at all.

1 · Choose a reference Cleanifier's human index, one click. Or your own FASTA — host genome, PhiX, vector, rRNA, a contaminating strain.
2 · Drop your FASTQ Reads are matched against the index by spaced seeds. Mates are decided together, so pairs stay intact.
3 · Two files out .unmatched is your cleaned sample. .matched is what was removed — so you can check the tool, not trust it.

1 · Reference

Drop the reference FASTA or click to browse
.fasta / .fa / .fna, optionally .gz — reads matching it go to one file, the rest to another

Spaced seeds tolerate substitutions better than contiguous k-mers of the same weight: on synthetic reads at 10% divergence, the Cleanifier seed recovered 14.6 points more matching reads than a contiguous k=29. They cost roughly 3× in throughput.

No index in memory.

2 · Samples

Drop your FASTQ files or click to browse
.fastq[.gz] / .fq[.gz] — _1/_2 and _R1/_R2 pairs are grouped automatically

For a pair, the worse mate decides both reads — splitting one without its mate would break mate pairing downstream.

Both sides are written. Each sample yields *.matched.fastq (reads found in the reference) and *.unmatched.fastq (the rest). Which one you keep is your call — the tool does not assume that matching the reference means contamination.

3 · Output folder

No folder selected.

Files are written straight to disk as they are produced. This avoids browser storage, whose quota would not hold a large FASTQ.

Compression happens in the same stream, so nothing extra is held in memory. It costs some throughput but roughly a third of the disk space.

idle

Log

ready.

Please cite

This tool reimplements the method described in the following work. If you use it, please cite:

Zentgraf, J., Schmitz, J.E., Rahmann, S. Cleanifier: Contamination removal from microbial sequences using spaced seeds of a human pangenome index. Bioinformatics (2025). doi.org/10.1093/bioinformatics/btaf632

If you use their pre-built index:

Schmitz, J.E., Zentgraf, J., & Rahmann, S. (2025). Human index for Cleanifier (0.1.0). Zenodo. doi.org/10.5281/zenodo.15639519