Fasta-O-Matic: a tool to sanity check and if needed reformat FASTA files

Abstract

As the sheer volume of bioinformatic sequence data increases, the only way to take advantage of this content is to more completely automate robust analysis workflows. Analysis bottlenecks are often mundane and overlooked processing steps. Idiosyncrasies in reading and/or writing bioinformatics file formats can halt or impair analysis workflows by interfering with the transfer of data from one informatics tools to another. Fasta-O-Matic automates handling of common but minor format issues that otherwise may halt pipelines. The need for automation must be balanced by the need for manual confirmation that any formatting error is actually minor rather than indicative of a corrupt data file. To that end Fasta-O-Matic reports any issues detected to the user with optionally color coded and quiet or verbose logs.

Fasta-O-Matic can be used as a general pre-processing tool in bioinformatics workflows (e.g. to automatically wrap FASTA files so that they can be read by BioPerl). It was also developed as a sanity check for bioinformatic core facilities that tend to repeat common analysis steps on FASTA files received from disparate sources. Fasta-O-Matic can be set with format requirements specific to downstream tools as a first step in a larger analysis workflow.

Fasta-O-Matic is available free of charge to academic and non-profit institutions at https://github.com/i5K-KINBRE-script-share/read-cleaning-format-conversion/tree/master/KSU_bioinfo_lab/fasta-o-matic.

Introduction

Sequence data can be stored as text with each letter representing a nucleic acid (DNA and RNA) or amino acid (protein). The linear nature of these molecules makes it natural to represent them as strings, finite sequences of characters. Although it has been argued that a graph, a network of edges connected by vertices, is a more accurate way to store genomic sequences because graphs allow the inclusion of alternate alleles and alternate possible assemblies (Jaffe 2012) all of the most common methods for storing sequences (FASTA, FASTQ, SAM/BAM) use a linear strings.

Other decisions about how to represent sequence data can be more arbitrary. For example, any character that is not used as a base or an amino acid can be used to indicate the beginning of a new sequence. Additionally text can be wrapped to limit the information content in any one line of a file. The advantage of wrapping text is that some programs can then be designed to work one line at time limiting the burden of each step (e.g. the program would never have to process an entire chromosome of sequence data in a single step). The disadvantage is that code must be slightly more complex to load an entire sequence record into the working memory.

FASTA file format specifications versus recommendations

FASTA file format requirements are very minimal (NCBI Accessed: 2015-08-06). Each sequence is preceded by a header/description line that begins with a aureplacedverbatimaa symbol. Sequence li