malign

malign performs a global anchored alignment between a reference sequence (in fasta format) and a collection of reads (in fastq format). It is designed to align bisulfite-converted reads from an amplicon to a reference sequence.

Installation

In order to run malign you need to install EMBOSS and have needle in the executable path. A convenient solution is to install via conda. First install miniconda (or anaconda) and then type the following:

conda config --add channels defaults
conda config --add channels bioconda
conda config --add channels conda-forge
conda config --set channel_priority strict
conda create --name emboss emboss
conda activate emboss

Global alignment is not computationally efficient: it is slow and uses a lot of memory. For this reason, the reference sequence should be a fragment just large enough to cover the amplion region.

Usage

The malign usage statement is reported if you don't give it the right arguments.

usage: malign [options] <fasta> <fastq>
options:
  -p <int>  percent identity minimum [90]
  -t <int>  threads [4]
  -w <int>  wrap [50]
  -d <path> create and keep a named working directory for temp files
  -m        methyl mode
              uses an asymmetric scoring matrix with C:T matching
              computes percent identity from A/G only
  -f        force rewrite over named working directory

Notes:

(1) By default, alignment is performed with a +1/-1 scoring matrix. in "methyl mode", the scoring matrix gives +1 to C:T matches in one direction and -1 in the other.

(2) Percent identity is calculated by ignoring gaps. In "methyl mode", percent identity is calculated from only the As and Gs in the reference sequence.

(3) To examine the temporary files, use the -d option to keep the working directory. This contains all of the intermediate FASTA files and pairwise alignments.

Example

malign -md build reference.fa reads.fq