include module type of struct include Biocaml_unix.Fasta end
FASTA files. The FASTA family of file formats has different incompatible descriptions (1, 2, 3, 4, etc.). Roughly FASTA files are in the format:
# comment
# comment
...
>description
sequence
>description
sequence
...
Comment lines are allowed at the top of the file. Usually comments start with a '#' but sometimes with a ';' character. The fmt
properties allow configuring which is allowed during parsing and printing.
Description lines begin with the '>' character. Various conventions are used for the content but there is no requirement. We simply return the string following the '>' character.
Sequences are most often a sequence of characters denoting nucleotides or amino acids, and thus an item
's sequence
field is set to a string. Sequences may span multiple lines.
However, sequence lines sometimes are used to provide quality scores, either as space separated integers or as ASCII encoded scores. To support the former case, we provide the sequence_to_int_list
function. For the latter case, see modules Phred_score
and Solexa_score
.
FASTA files are used to provide both short sequences and very big sequences, e.g. a genome. In the latter case, the main API of this module, which returns each sequence as an in-memory string, might be too costly. Consider using instead the read0
function which does not merge multiple sequence lines into one string. This API is slightly more difficult to use but perhaps a worthwhile trade-off.
Some FASTA files include very large sequences on a single line. This is discouraged and not well supported by this module. Functions in this module require memory proportional to the length of a line. Thus, a whole chromosomal sequence on a single line will consume a large amount of memory. This might not be a problem given the RAM on most computers.
Format Specifiers:
Variations in the format are controlled by the following settings, all of which have a default value. These properties are combined into the fmt
type for convenience and the defaults into default_fmt
.
allow_sharp_comments
: Allow comment lines beginning with a '#' character. Default: true.
allow_semicolon_comments
: Allow comment lines beginning with a ';' character. Default: false.
Setting both allow_sharp_comments
and allow_semicolon_comments
allows both. Setting both to false disallows comment lines.
allow_empty_lines
: Allow lines with only whitespace anywhere in the file. Default: false.
comments_only_at_top
: Allow comments only at the top of the file. If false, comment lines can occur anywhere but only the ones at the top are returned. The rest are ignored. Default: true.
max_line_length
: Require sequence lines to be shorter than given length. None means there is no restriction. Note this does not restrict the length of an item
's sequence
field because this can span multiple lines. Default: None.
alphabet
: Require sequence characters to be at most those in given string. None means any character is allowed. Default: None.
A header is a list of comment lines.
type fmt = Biocaml_unix.Fasta.fmt = {
allow_empty_lines : bool;
max_line_length : int option;
alphabet : string option;
}
Low-level Parsing
type item0 = private [<
| `Empty_line
| `Description of string
| `Partial_sequence of string
]
An item0
is more raw than item
. It is useful for parsing files with large sequences because you get the sequence in smaller pieces.
`Comment _
- Single comment line without the final newline. Initial comment char is retained.
`Empty_line
- Got a line with only whitespace characters. The contents are not provided.
`Description _
- Single description line without the initial '>' nor final newline.
`Partial_sequence _
- Multiple sequential partial sequences comprise the sequence of a single item
.