Newmap

Introduction

Newmap is a software package that efficiently identifies uniquely mappable regions of any genome. It accomplishes this task by outputting read lengths at every position that are unique to that genome. From the range of unique read lengths produced, the single-read mappability and the multi-read mappability for a specific read length can be generated.

Newmap can search for unique k-mer/read lengths on specific values, or entire continuous ranges using a binary search method allowing for finding the minimum possible unique k-mer/read length.

Newmap requires a CPU that supports the AVX2 instruction set.

OpenMP is required for parallel processing.

Documentation

The latest for Newmap is available on Read the Docs.

All commands have a --help option to provide additional usage information.

Quick start

Installation

Python Package Index (PyPI)

pip install newmap

Bioconda

conda install newmap

Usage

1. Create an index for a genome

newmap index genome.fa

By default this will create a genome.awfmi file in the current directory.

2. Find the minimum unique k-mer lengths for the genome using the index

Searching the entire genome, using 20 threads, printing status information, and searching lengths ranging from 20 to 200 bp:

newmap search --verbose --num-threads=20 --search-range=20:200 --output-directory=unique_lengths genome.fa

This will create *.unique.uint8 files (one for each sequence ID) in the unique_lengths directory.

3. Convert the unique lengths to mappability tracks

To output single-read and multi-read mappability for a 24 bp read length:

newmap track --single-read=24.bed --multi-read=24.wig 24 unique_lengths/*.unique.uint8

For both single-read and multi-read mappability, this will generate a single file that contains the mappability for all sequences listed in the unique_lengths directory. The resulting BED file will be the single read mappability, and the WIG file will be the multi-read mappability.

Credits

Newmap is a reimplementation of the output of Umap. Umap was developed by Mehran Karimzadeh. The repository for that implmemention is found at https://www.github.com/hoffmangroup/umap. Umap in turn was originally developed by Anshul Kundaje and was written in MATLAB. The original repository is available https://sites.google.com/site/anshulkundaje/projects/mappability.

This project uses the excellent AvxWindowFMIndex library. Read their published article here (https://doi.org/10.1186/s13015-021-00204-6)