Convert various sequence formats to FASTA

You may wonder why this tool even exists. Well, I tried to do the right thing and use established tools like readseq and seqret from EMBOSS, but they both mangled IDs containing | or . characters, and there is no way to fix this behaviour. This resulted in inconsitences between my .gbk and .fna versions of files in my pipelines.

Then you may wonder why I didn't use Bioperl or Biopython. Well they are heavyweight libraries, and actually very slow at parsing Genbank files. This script uses only core Perl modules, has no other dependencies, and runs very quickly.

It supports the following input formats:

1. Genbank flat file, typically .gb, .gbk, .gbff (starts with LOCUS)
2. EMBL flat file, typically .embl, (starts with ID)
3. GFF with sequence, typically .gff, .gff3 (starts with ##gff)
4. FASTA DNA, typically .fasta, .fa, .fna, .ffn (starts with >)
5. FASTQ DNA, typically .fastq, .fq (starts with @)
6. CLUSTAL alignments, typically .clw, .clu (starts with CLUSTAL or MUSCLE)
7. STOCKHOLM alignments, typically .sth (starts with # STOCKHOLM)
8. GFA assembly graph, typically .gfa (starts with ^[A-Z]\t)

Files may be compressed with:

1. gzip, typically .gz
2. bzip2, typically .bz2
3. zip, typically .zip

any2fasta has no dependencies except Perl 5.10 or higher. It only uses core modules, so no CPAN needed.

% cd /usr/local/bin  # choose a folder in your $PATH
% wget https://raw.githubusercontent.com/tseemann/any2fasta/master/any2fasta
% chmod +x any2fasta

% brew install brewsci/bio/any2fasta

% conda install -c bioconda any2fasta

% git clone https://github.com/tseemann/any2fasta.git
% cp any2fasta/any2fasta /usr/local/bin # choose a folder in your $PATH

% ./any2fasta -v
any2fasta 0.2.2

% ./any2fasta -h
NAME
  any2fasta 0.4.2
SYNOPSIS
  Convert various sequence formats into FASTA
USAGE
  any2fasta [options] file.{gb,fa,fq,gff,gfa,clw,sth}[.gz,bz2,zip] > output.fasta
OPTIONS
  -h       Print this help
  -v       Print version and exit
  -q       No output while running, only errors
  -n       Replace ambiguous IUPAC letters with 'N'
  -l       Lowercase the sequence
  -u       Uppercase the sequence
END

% any2fasta ref.gbk > ref.fna

% any2fasta in.fasta > out.fasta  # should behave like "cat"

% any2fasta prokka.gff > prokka.fna  # only if GFF has FASTA appended

% any2fasta - < file.gb > file.fasta  # '-' means stdin

% anyfasta genes.gff.gz > genes.ffn  # automatically decompresses

% any2fasta 1.gb 2.fa.gz 3.gff.bz2 - > out.fa  # multiple files and stdin

% any2fasta R1.fq.gz | bzip2 > R1.fa.bz2  # 'seqtk seq -A' is much faster

% any2fasta -q 23S.clw > 23S.aln  # gaps '-' will be preserved

% any2fasta pfam4321.sth > pfam4321.aln  # '.' gaps will become '-'

• -n replaces any characters that aren't A,C,G,T with N (gaps preserved)
• -l will lowercase all the letters
• -u will uppercase all the letters
• -q will prevent logging messages being printed

Submit feedback to the Issue Tracker

GPL v3

Torsten Seemann