Gaussian Direct Coupling Analysis for protein contacts predicion

Overview

This is the code which accompanies the paper "Fast and accurate multivariate Gaussian modeling of protein families: Predicting residue contacts and protein-interaction partners" by Carlo Baldassi, Marco Zamparo, Christoph Feinauer, Andrea Procaccini, Riccardo Zecchina, Martin Weigt and Andrea Pagnani, (2014) PLoS ONE 9(3): e92721. doi:10.1371/journal.pone.0092721

See also this Wikipedia article for a general overview of the Direct Coupling Analysis technique.

This code is released under the GPL version 3 (or later) license; see the LICENSE.md file for details.

The code is written in Julia and requires julia version 0.6 or later; it provides a function which reads a multiple sequence alignment (in FASTA format) and returns a ranking of all pairs of residue positions in the aligned amino-acid sequences.

If you use the code in your research, please cite the abovementioned paper and the following DOI:

Installation

Install the package by giving these commands:

julia> using Pkg # only in Julia 0.7 or later

julia> Pkg.clone("https://github.com/carlobaldassi/GaussDCA.jl")

All dependencies will be downloaded and installed automatically.

In Julia 0.6, the command Pkg.upadte() will fetch the latest changes from this repository.

In Julia 0.7 and later, however, if you want to update you need to do so explicitly from the package directory using git. One way to do that is as such:

julia> using Pkg

julia> cd(joinpath(Pkg.devdir(), "GaussDCA"))

shell> git pull origin master

Note that the last line is given from the shell prompt, which you can access by pressing the ; key.

Usage

To load the code, just type using GaussDCA.

This software provides one main function, gDCA(filname::String, ...). This function takes the name of a (possibly gzipped) FASTA file, and returns a predicted contact ranking, in the form of a Vector of triples, each triple containing two indices i and j (with i < j) and a score. The indices start counting from 1, and denote pair of residue positions in the given alignment; pairs which are separated by less than a given number of residues (by default 5) are filtered out. The triples are sorted by score in descending order, such that predicted contacts should come up on top.

For convenience, a utility function is also provided, printrank(output, R), which prints the result of gDCA either in a file or to a stream, given as first argument. If the first argument output is omitted, the standard terminal output will be used.

The gDCA function takes some additional, optional keyword arguments:

pseudocount: the value of the pseudo-count parameter, between 0 and 1. the default is 0.8, which gives good results when the Frobenius norm score is used (see below); a good value for the Direct Information score is 0.2.
theta: the value of the similarity threshold. By default it is :auto, which means it will be automatically computed (this takes additional time); otherwise, a real value between 0 and 1 can be given.
max_gap_fraction: maximum fraction of gap symbols in a sequence; sequences which exceed this threshold are discarded. The default value is 0.9.
score: the scoring function to use. There are two possibilities, :DI for the Direct Information, and :frob for the Frobenius norm. The default is :frob. (Note the leading colon: this argument is passed as a symbol).
min_separation: the minimum separation between residues in the output ranking. Must be >= 1. The default is 5.

The code will be parallelized if more than one julia worker (as obtained by the nworkers() function) is available. Multiple workers can be created either by launching julia with the -p option from the command line, or by using the addprocs function (note that since julia 0.7 you will need to execute using Distributed before you can call addprocs). See also the "Additional thechnical notes" section at the end of this document.

Examples

Here is a basic usage example, assuming an alignment in FASTA format is found in the file "alignment.fasta.gz":

julia> using GaussDCA

julia> FNR = gDCA("alignment.fasta.gz");

julia> printrank("results_FN.txt", FNR)

The above uses the Frobenius norm ranking with default parameters. This is how to get the Direct Information ranking instead:

julia> DIR = gDCA("alignment.fasta.gz", pseudocount = 0.2, score = :DI);

julia> printrank("results_DI.txt", DIR)

Additional technical details

The parallelization can be forcefully disabled even in presence of extra workers, by setting the environment variable PARALLEL_GDCA to false before loading the GaussDCA module.

When using workers, and using either OpenBLAS - which is the default - or MKL as the BLAS backend, the default behaviour is to disable threading in BLAS libraries. In this case, i.e. when many workers are found and parallelization is not manually disabled, the gDCA function overrides the default julia behaviour and sets the number of threads to match the number of workers (except when running the parallel portions of the code). It then resets the number of threads to 1 when finished. The number of cores used in the non-parallel portions of the code can be explicitly controlled by the user via the OMP_NUM_THREADS environment variable.

pagnani / GaussDCA.jl