view release on metacpan or  search on metacpan

bin/classify-ali.pl  view on Meta::CPAN

=pod

=head1 NAME

classify-ali.pl - Classify ALI files based on taxonomic filters

=head1 VERSION

version 0.252040

=head1 USAGE

    classify-ali.pl <indirs> --config=<file> --taxdir=<dir>
        [optional arguments]

=head1 REQUIRED ARGUMENTS

=over

=item <indirs>

Path to input directories containing ALI files [repeatable argument].

=for Euclid: indirs.type: string
    repeatable

=item --config=<file>

Path to the configuration file specifying the classifier details.

In principle, several configuration file formats are available: XML, JSON,
YAML. However, this program was designed with YAML in mind.

The configuration file defines different 'categories'. The order of
definition is relevant. Hence, if an ALI matches more than one category, it
is classified according to the first one that was defined. Each category has
a 'label' that is used to create the corresponding subdirectory for sorting
ALI files.

A category is characterized by one or more 'criteria'. To match a category,
an ALI must satisfy all criteria. Criteria are thus linked by logical ANDs
(and their order of definition is irrelevant).

Each criterion has a 'tax_filter' describing its taxonomic requirements. Wanted
taxa are to be prefixed by a C<+> symbol, whereas unwanted taxa are to be
prefixed by a C<-> symbol. Wanted taxa are linked by logical ORs while unwanted
taxa are linked by logical ANDs. Unwanted taxa should not include wanted taxa
because the former ones take precedence over the latter ones. In contrast the
opposite helps fine-tuning the tax_filter.

Criteria may also have a 'min_seq_count' and a 'max_seq_count' arguments.
These respectively specify the minimum and maximum number of sequences that
must pass the tax_filter for the ALI to match the criterion. Minimum
defaults to 1, while there is no upper bound by default.

Other conditions are available: 'min_org_count' and 'max_org_count' deal with
organisms instead of sequences, whereas 'min_copy_mean' and 'max_copy_mean'
allow bounding the mean number of gene copies per organism. Moreover, there
exist 'perc' variants ('min_seq_perc', 'max_seq_perc', 'min_org_perc',
'max_org_perc') that count in percents (either of total sequences or total
organisms). All default to no bound.

An example YAML file follows:

    categories:
    - label: strict
      description: strict species sampling
      criteria:
      - tax_filter: [ +Latimeria ]
        min_seq_count: 1
        max_seq_count:
        min_org_count:
        max_org_count:
        min_copy_mean:
        max_copy_mean:
        min_seq_perc:
        max_seq_perc:
        min_org_perc:
        max_org_perc:
      - tax_filter: [ +Protopterus ]
      # min_seq_count defaults to 1
      # max_seq_count defaults to no upper bound
      # all other also default to no bound
      - tax_filter: [ +Danio, +Oreochromis ]
      - tax_filter: [ +Xenopus ]
      - tax_filter: [ +Anolis, +Gallus, +Meleagris, +Taeniopygia ]
      - tax_filter: [ +Mammalia ]
    - label: loose
      description: loose species sampling
      criteria:
      - tax_filter: [ +Latimeria ]
      - tax_filter: [ +Protopterus ]
      - tax_filter: [ +Danio, +Oreochromis ]
      - tax_filter: [ +Amphibia, +Amniota ]

=for Euclid: file.type: readable

=item --taxdir=<dir>

Path to local mirror of the NCBI Taxonomy database.

=for Euclid: dir.type: string

=back

=head1 OPTIONAL ARGUMENTS

=over

=item --version

=item --usage

=item --help

=item --man

Print the usual program information

=back