Ctrl+K

Biological Data Formats

8.1. Biological Data Formats#

Editor Wars
Read the original at XKCD

Oh dear, this is a mess. Not this document, but the “multitude” of data format standards in Bioinformatics. Bioinformatics is replete with different variants on basic formats. Not all standards are well documented and support for them is … well, I hope your experience is better than mine. The fact that data formats can be such a problem is a great argument for knowing how to code, since that skill gives you the ability to handle the different format variants that no existing tool handles.

The above aside, I will briefly describe and link [1] to the canonical description of a data format for the main types of data. I’ve grouped these into formats related to the major biological data types of sequences, genomic features, and phylogenetic trees.

8.1.1. Sequences and alignments#

8.1.1.1. fasta#

The most commonly, and probably the easiest to parse, format is that of fasta. There are too many variants of this basic format, so I just link to the Wikipedia entry and then present an example below.:

>seqlabel 1 line
ACCGGTGA
AAG
>seqlabel2
AGGCG

While this is listed as a sequence format it’s often used to include multiple sequences, which may be aligned.

8.1.1.2. genbank#

GenBank is a web portal to a multitude of valuable databases and bioinformatic tools. They also define a number of data file formats and, in particular, the genbank format for sequence data. This is format includes at a minimum extensive meta-data about a database entry. They do not always contain sequence, but will at least reference identifiers of related sequence records in their database. Below is a sample record from GenBank (which I’ve edited for brevity)

LOCUS       AF165912                5485 bp    DNA     linear   PLN 29-JUL-1999
DEFINITION  Arabidopsis thaliana CTP:phosphocholine cytidylyltransferase (CCT)
            gene, complete cds.
ACCESSION   AF165912
VERSION     AF165912.1
KEYWORDS    .
SOURCE      Arabidopsis thaliana (thale cress)
  ORGANISM  Arabidopsis thaliana
            Eukaryota; Viridiplantae; Streptophyta; Embryophyta; Tracheophyta;
            Spermatophyta; Magnoliophyta; eudicotyledons; Gunneridae;
            Pentapetalae; rosids; malvids; Brassicales; Brassicaceae;
            Camelineae; Arabidopsis.
REFERENCE   1  (bases 1 to 5485)
  AUTHORS   Choi,Y.H., Choi,S.B. and Cho,S.H.
  TITLE     Structure of a CTP:Phosphocholine Cytidylyltransferase Gene from
            Arabidopsis thaliana
  JOURNAL   Unpublished
REFERENCE   2  (bases 1 to 5485)
  AUTHORS   Choi,Y.H., Choi,S.B. and Cho,S.H.
  TITLE     Direct Submission
  JOURNAL   Submitted (06-JUL-1999) Biology, Inha University, Yonghyon-Dong
            253, Inchon 402-751, Korea
FEATURES             Location/Qualifiers
     source          1..5485
                     /organism="Arabidopsis thaliana"
                     /mol_type="genomic DNA"
                     /db_xref="taxon:3702"
                     /ecotype="Col-0"
     gene            1..4637
                     /gene="CCT"
     regulatory      1..1602
                     /regulatory_class="promoter"
                     /gene="CCT"
<features removed for brevity>
ORIGIN
        1 ccagaatggt tactatggac atccgccaac catacaagct atggtgaaat gctttatcta
       61 tctcattttt agtttcaaag cttttgttat aacacatgca aatccatatc cgtaaccaat
      121 atccaatcgc ttgacatagt ctgatgaagt ttttggtagt taagataaag ctcgagactg

8.1.1.3. gff – general feature format#

The GFF format, although widely used, has fragmented into multiple incompatible dialects.

—Lincoln Stein, in the prologue to defining gff3.

Sounds like the cartoon at the top doesn’t it! Anyway, it is widespread and an important format for storing information about sequence annotations [2], so here’s the canonical gff3 definition. This is a tab delimited format with 9 distinct fields. It’s the last field, attributes, that proves to be the most difficult to parse. Below is a small sample of a file posted on the definition page.

0  ##gff-version 3.1.25
1  ##sequence-region ctg123 1 1497228
2  ctg123 . gene            1000  9000  .  +  .  ID=gene00001;Name=EDEN
3  ctg123 . TF_binding_site 1000  1012  .  +  .  ID=tfbs00001;Parent=gene00001

8.1.2. newick format for phylogenetic trees#

This is the most widespread text format for dsitributing phylogenetic trees. Clades of lineages are denoted by () and separate by , and can be grouped into subclades. Branch lengths are indicated by numbers after a colon character. There is some funky behaviour around dealing with spaces in tip names, they are often represented in the name as an underscore character ("_"). If you can, avoid any issues by not having spaces or underscores in names. Here’s a sample.

((Human:0.006062440217780064,Chimpanzee:0.003020541234140796):0.09488527928524751,((Mouse:0.06659142318491332,Rat:0.05783486638653178):0.17244926332734278,Wombat:0.4522900123679113):0.0424545337445269,Horse:0.05802695948476483);

Contents