The demultiplex.py tool is used to separate sequencing reads into their corresponding samples based on the presence of internal barcodes. In high-throughput sequencing experiments, multiple samples are often pooled together into a single sequencing run to save costs. Each sample is tagged with a short, unique barcode sequence so that later, during analysis, reads can be assigned back to the correct sample. The purpose of this tool is to read raw FASTQ files containing all mixed samples and generate new FASTQ files for each individual sample, plus a summary of the process.

The program scans the input forward (and optionally reverse) reads for barcode sequences that match those provided in a barcode mapping file. Depending on user settings, it can allow a certain number of mismatches in barcode recognition and can check barcodes at the beginning of the forward read, the end of the reverse read, or both. Reads that match a barcode are written into per-sample FASTQ files, packaged into a single archive, while unassigned reads are collected separately. Alongside the demultiplexed reads, the tool produces a summary table showing how many reads were correctly assigned to each sample and how many were excluded, ensuring quality control and transparency in the demultiplexing step.