KRIS Read Complexity Filter

A high-performance tool for filtering single-end or paired-end FASTQ reads based on k-mer entropy using the KRIS index.

Key Features

• Automatically determines k-mer size from read length
• Filters reads below a specified KRIS entropy threshold
• Handles plain and gzipped FASTQ
• Outputs summary statistics and SVG visualizations
• Docker support for reproducibility

Installation

Pip (local use)

pip install .

Docker

docker build -t kris-read-complexity-filter .

Usage

Command-line interface

kris-read-complexity-filter \
  --in1 input_R1.fastq.gz \
  --in2 input_R2.fastq.gz \
  --out1 output_R1.filtered.fastq \
  --out2 output_R2.filtered.fastq \
  --threshold 1.8 \
  --report-dir reports/

If single-end, omit --in2 and --out2.

Output

• Filtered FASTQ files
• summary.json: total reads, filtered count, retained reads, filter rate
• kris_score_histogram.svg: histogram of KRIS scores
• kris_score_boxplot.svg: boxplot comparing retained vs total scores

Example

Filter paired-end reads with a minimum KRIS index of 2.0:

kris-read-complexity-filter \
  --in1 sample_R1.fastq.gz \
  --in2 sample_R2.fastq.gz \
  --out1 clean_R1.fastq \
  --out2 clean_R2.fastq \
  --threshold 2.0