lh3 / pre-pe

Preprocessing paired-end reads produced with experiment-specific protocols

Pre-pe is a set of tools to preprocess paired-end reads. They trim generic sequencing adapters, clip experiement specific adapters, identify inline barcodes and merge overlapping ends. Each tool in pre-pe has very similar functionality but for different data types (see below). The general command line to run pre-pe looks like:

seqtk mergepe read1.fq.gz read2.fq.gz | pre-adna - | gzip > pe-se-mixed.fq.gz

Here, the seqtk command line generates an interleaved FASTQ stream. You can skip this step if your FASTQs are already interleaved. By default, paired-end and merged single-end reads are merged into a single stream. You can use bwa-mem to directly map the output with

pre-meta interleaved.fq.gz | bwa mem -p ref.fa - | gzip > output.sam.gz

Most other short-read mappers don't have this functionality.

Pre-pe consists of the following tools:

• pre-lianti for single-cell whole-genome sequencing data produced with the LIANTI protocol. It is the first tool in this series and also available from my lianti repo.

• pre-adna for ancient data produced with the Reich Lab protocol. This program is also available from the adna repo. It has been used to process hundreds of ancient full genomes from the Reich lab.

• pre-dip-c for single-cell Hi-C data produced with the Dip-C protocol. It was modified from pre-lianti by Longzhi Tan.

• pre-meta for single-cell genomic data produced with the META protocol. It is similar to pre-dip-c except that it additionally checks ambiguous end merges to greatly reduce artifactual deletions.

Of these tools pre-adna may be of general interest.