demultiplex utility for paired-end eCLIP raw fastq files (process eCLIP barcodes and ramdomers) See: https://github.com/YeoLab/eclipdemux for full code and documentation Usage: eclipdemux --dataset DATASET_ID --metrics METRICS_FILE --fastq_1 READ_1 --fastq_2 READ_2 --expectedbarcodeida BARCODE_A --expectedbarcodeidb BARCODE_B --barcodesfile BARCODES_FASTA --length LENGTH

This tool wraps parsebarcodes.sh.

We have observed occasional double ligation events on the 5’ end of Read1, and we have found that to fix this requires we run cutadapt twice. Additionally, because two adapters are used for each library (to ensure proper balancing on the Illumina sequencer), two separate barcodes may be ligated to the same Read1 5’ end (often with 5’ truncations). To fix this we split the barcodes up into 15bp chunks so that cutadapt is able to deconvolute barcode adapters properly (as by default it will not find adapters missing the first N bases of the adapter sequence)

parsebarcodes.sh writes the following files: trimfirst_overlap_length.txt : file that always contains \"1\" trimagain_overlap_length.txt : file that contains max((length of longest barcode - 2),5) g_adapters_default.fasta : empty file (to be fed to cutadapt properly) a_adapters_default.fasta : empty file (to be fed to cutadapt properly) g_adapters.fasta : fasta file containing sequences to be trimmed via cutadapt -g flag a_adapters.fasta : fasta file containing sequences to be trimmed via cutadapt -a flag A_adapters.fasta : fasta file containing sequences to be trimmed via cutadapt -A flag

Usage: parsebarcodes.sh <randommer_length> <barcodes_fasta> <barcode_A> <barcode_B>