MuffinEC 2.0b
Contact: asalic@posgrado.upv.es
--------
Overview
--------
MuffinEc is an error correction software for NGS datasets. It can handle multiple technologies by its ability of correcting mismatches as well as insertions and deletions. It uses a greedy approach where it groups reads having a certain number of kmers in common. Afterwards, MuffinEc refines each group using a custom vertion of the Smith-Waterman algorithm, generating multiple sequence alignments structures. Finally, it corrects each of these structures using a "base holding the majority wins" approach on each column.
-------------------
System Requirements
-------------------
The software was built and tested under Debian and Ubuntu Linux x64. The program should compile and run under Mac OS X like it does under Linux. MuffinEc requires GCC version is 4.7.2 or above with support for OpenMP 3.0 or higher and C++11. We advise the user to utilize a 64 bit system because we are heavily making use of 64bit integers through our code.
-----------
Compilation
-----------
- <version> denotes the version of the program (major followed by minor), e.g. "muffinec1_0.tar.gz"
Linux (and Mac) Users:
- download the archive file "muffinec<version>.tar.gz"
- uncompress it in the current directory:
tar -xvf muffinec<version>.tar.gz
- enter the newly created directory:
cd muffinec<version>/
- run make to compile the code
make
- finally, an executable called "muffinec" will be created in "build" directory
-----
Usage
-----
MuffinEC allows the user to fine-tune many of its parameters. This way it can adapt to a variety of datasets, without recompilation. There are only three mandatory parameters, namely the technology used, the input file path and the output file path. Bellow it's a short list with all available parameters.
MuffinEC creates the k-table in a separate process than the actual correction. As a result, the first argument must be the name of the step (either "ktbl" or "correct").
Programs:
****************
muffinec ktbl
****************
Mandatory parameters:
-a, --fasta: The full path of the FASTA file containing the reads to be corrected; Must be set if -q is not (exclusive parameter)
-q, --fastq: The full path of the FASTQ file containing the reads to be corrected; Must be set if -a is not (exclusive parameter)
--454: The input data is from a Roche 454; Must be set if none of --illumina, --pac, --ion or --generic are set (exclusive parameter)
--illumina: The input data is from Illumina technologies; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--pac: The input data is from Pacific Biosiences; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--ion: The input data is from Ion Torrent; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--generic: Use this option if you don't want to set/don't know the source technology; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
Optional Parameters:
-k, --kmerlen: The length of the kmer used to do the coverage and to create the neighborhood [15]
--minsn: Minimum size of a neighborhood, created by the greedy grouping method and which will be further processed by the correction mechanism [3]
--trimq: Value marking the minimum accepted quality score for a base at the ends of the reads such that the position won't get axed [43]
--kmerq: Minimum accepted quality score for each base in a kmer [0]
-s, --hashk: The hashing method used to store kmers; The possible options are: umap and vector [umap]
*******************
muffinec correct
*******************
Mandatory parameters:
-a, --fasta: The full path of the FASTA file containing the reads to be corrected; Must be set if -q is not (exclusive parameter)
-q, --fastq: The full path of the FASTQ file containing the reads to be corrected; Must be set if -a is not (exclusive parameter)
-o, --output: The name of the output file containing the corrected reads
--454: The input data is from a Roche 454; Must be set if none of --illumina, --pac, --ion or --generic are set (exclusive parameter)
--illumina: The input data is from Illumina technologies; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--pac: The input data is from Pacific Biosiences; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--ion: The input data is from Ion Torrent; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
--generic: Use this option if you don't want to set/don't know the source technology; Must be set if none of --454, --pac, --ion or --generic are set (exclusive parameter)
Optional Parameters:
-p, --threads: Number of OpenMP threads used by the program [1]
-m, --pmj: The percent of reads which must have the same base on a column of the consensus [0.51]
--almis: SW Aligner mismatch score [-1]
--almat: SW Aligner match score [2]
--algap: SW Aligner gap opening score [-3]
--algapext: SW Aligner gap extending score [-1]
--pswerr: The max number of errors accepted by the Smith Waterman algorithm when comparing the distance between a read and a subgroup; if the total number of differences between the read and the subgroup are greater than the limit set by this parameter the read won't be added to the subgroup [0.1]
--maxsn: Maximum size of a neighborhood, created by the grouping method and which will be further processed by the correction mechanism [500]
--perrov: The percentage of errors found in the overlap region of two reads compared by the fast gapped kmer algorithm [0.02]
--pmincomk: Percentage of overlapping for two reads, calculated from the smaller between the two [0.5]
--pminovsw: Minimum SW determined overlap percentage for a read over consensus such that it will be considered a part of the consensus [0.0]
--gencon: Add this flag to generate the contigs for the consensus [false]
--allowIndelsArg: Add this flag to allow the SW algorithm to use indels when aligning two reads [false]
-------
Example
-------
To correct a 454 dataset, one must first generate the k-mer table and then start the correction.
./muffinec ktbl --454 -a test.fa
./muffinec correct -p 6 -a test.fa -o test_corr.fa
----------------------------
License and Third-Party Libs
----------------------------
MuffinEc uses TCLAP to parse the input parameters. The aforementioned library is released under The MIT License (http://opensource.org/licenses/mit-license.php). MuffinEc is licensed under LGPL3.
-----------
How to Cite
-----------
Alic, A. S., Tomas, A., Medina, I., & Blanquer, I. (2016). MuffinEc: Error correction for de Novo assembly via greedy partitioning and sequence alignment. Information Sciences, 329, 206-219.
