diverse-seq 2024.9.2a1

diverse_seq: a tool for sampling diverse biological sequences

diverse_seq identifies the most diverse biological sequences from a collection

diverse_seq provides tools for selecting a representative subset of sequences from a larger collection. It is an alignment-free method which scales linearly with the number of sequences. It identifies the subset of sequences that maximize diversity as measured using Jensen-Shannon divergence. diverse_seq provides a command-line tool (dvs) and plugins to the Cogent3 app system (prefixed by dvs_) allowing users to embed code in their own scripts. The command-line tools can be run in parallel.

dvs prep: preparing the sequence data

Convert sequence data into a more efficient format for the diversity assessment. This must be done before running either the nmost or max commands.

CLI options for dvs prep

Usage: dvs prep [OPTIONS]

  Writes processed sequences to a <HDF5 file>.dvseqs.

Options:
  -s, --seqdir PATH        directory containing sequence files  [required]
  -sf, --suffix TEXT       sequence file suffix  [default: fa]
  -o, --outpath PATH       write processed seqs to this filename  [required]
  -np, --numprocs INTEGER  number of processes  [default: 1]
  -F, --force_overwrite    Overwrite existing file if it exists
  -m, --moltype [dna|rna]  Molecular type of sequences, defaults to DNA
                           [default: dna]
  -L, --limit INTEGER      number of sequences to process
  -hp, --hide_progress     hide progress bars
  --help                   Show this message and exit.

dvs nmost: select the n-most diverse sequences

Selects the n sequences that maximise the total JSD. We recommend using nmost for large datasets.

Note A fuller explanation is coming soon!

Options for command line dvs nmost

Usage: dvs nmost [OPTIONS]

  Identify n seqs that maximise average delta JSD

Options:
  -s, --seqfile PATH       path to .dvtgseqs file  [required]
  -o, --outpath PATH       the input string will be cast to Path instance
  -n, --number INTEGER     number of seqs in divergent set  [required]
  -k INTEGER               k-mer size  [default: 6]
  -i, --include TEXT       seqnames to include in divergent set
  -np, --numprocs INTEGER  number of processes  [default: 1]
  -L, --limit INTEGER      number of sequences to process
  -v, --verbose            is an integer indicating number of cl occurrences
                           [default: 0]
  -hp, --hide_progress     hide progress bars
  --help                   Show this message and exit.

Options for cogent3 app dvs_nmost

The dvs nmost is also available as the cogent3 app dvs_nmost. The result of using cogent3.app_help("dvs_nmost") is shown below.

Overview
--------
select the n-most diverse seqs from a sequence collection

Options for making the app
--------------------------
dvs_nmost_app = get_app(
    'dvs_nmost',
    n=10,
    moltype='dna',
    include=None,
    k=6,
    seed=None,
)

Parameters
----------
n
    the number of divergent sequences
moltype
    molecular type of the sequences
k
    k-mer size
include
    sequence names to include in the final result
seed
    random number seed

Notes
-----
If called with an alignment, the ungapped sequences are used.
The order of the sequences is randomised. If include is not None, the
named sequences are added to the final result.

Input type
----------
SequenceCollection, ArrayAlignment, Alignment

Output type
-----------
SequenceCollection, ArrayAlignment, Alignment

dvs max: maximise variance in the selected sequences

The result of the max command is typically a set that are modestly more diverse than that from nmost.

Note A fuller explanation is coming soon!

Options for command line dvs max

Usage: dvs max [OPTIONS]

  Identify the seqs that maximise average delta JSD

Options:
  -s, --seqfile PATH       path to .dvtgseqs file  [required]
  -o, --outpath PATH       the input string will be cast to Path instance
  -z, --min_size INTEGER   minimum size of divergent set  [default: 7]
  -zp, --max_size INTEGER  maximum size of divergent set
  -k INTEGER               k-mer size  [default: 6]
  -st, --stat [stdev|cov]  statistic to maximise  [default: stdev]
  -i, --include TEXT       seqnames to include in divergent set
  -np, --numprocs INTEGER  number of processes  [default: 1]
  -L, --limit INTEGER      number of sequences to process
  -T, --test_run           reduce number of paths and size of query seqs
  -v, --verbose            is an integer indicating number of cl occurrences
                           [default: 0]
  -hp, --hide_progress     hide progress bars
  --help                   Show this message and exit.

Options for cogent3 app dvs_max

The dvs max is also available as the cogent3 app dvs_max.

Overview
--------
select the maximally divergent seqs from a sequence collection

Options for making the app
--------------------------
dvs_max_app = get_app(
    'dvs_max',
    min_size=5,
    max_size=30,
    stat='stdev',
    moltype='dna',
    include=None,
    k=6,
    seed=None,
)

Parameters
----------
min_size
    minimum size of the divergent set
max_size
    the maximum size of the divergent set
stat
    either stdev or cov, which represent the statistics
    std(delta_jsd) and cov(delta_jsd) respectively
moltype
    molecular type of the sequences
include
    sequence names to include in the final result
k
    k-mer size
seed
    random number seed

Notes
-----
If called with an alignment, the ungapped sequences are used.
The order of the sequences is randomised. If include is not None, the
named sequences are added to the final result.

Input type
----------
SequenceCollection, ArrayAlignment, Alignment

Output type
-----------
SequenceCollection, ArrayAlignment, Alignment