TRACS

TRACS provides robust estimates of pairwise transmission distances from single isolate, multi-strain and metagenomic samples. It uses an empirical Bayes approach to account for variable sequence coverage and aligns to multiple reference genomes to estimate a lower bound for both the SNP distance and the number of intermediate hosts separating two samples.

Note: TRACS is not intended to estimate very large SNP distances

Documentation

TRACS is currently under development and frequent backwards incompatible changes may be made.

Documentation for TRACS can be found here

Installation

Conda

TRACS can be installed using conda by running

conda install bioconda::tracs

Note: Conda and container builds are not guaranteed to be portable across all CPU generations due to microarchitecture differences. If you run into a SIGILL or Illegal instruction (core dumped) error when using the bioconda package or biocontainers image, it is likely because your CPU does not support the instructions the binary was built with. Building from source using pip can solve this problem.

Manual

TRACS is a python package and can be installed easily using pip.

pip3 install git+https://github.com/gtonkinhill/tracs

By default, building from source uses your compiler's default baseline architecture for maximum portability. If you want to optimise the build for your current CPU to obtain maximum performance, you can set the TRACS_MARCH environment variable before installing:

TRACS_MARCH=native pip3 install git+https://github.com/gtonkinhill/tracs

This is all that is needed for the pairwise distance and clustering commands. To generate alignments you will also need to install samtools, minimap2 and htsbox.