tkp-finder

tkp-finder is a CLI tool to discover and annotate tandem protein kinases.

It's based on lXtractor -- a general-purpose library for data mining from sequences and structures. The latter is under active development, so bugs are possible.

Table of Contents

• Installation
• License
• Usage

Installation

pip install tkp-finder

License

tkp-finder is distributed under the terms of the MIT license.

Usage

The installation should make the script tkp-finder globally available. The interface has two commands:

The setup command will download and prepare HMM models for annotation.

→ tkp-finder setup --help

Usage: tkp-finder setup [OPTIONS]

  Command to initialize the HMM data needed for TKPs' annotation.

Options:
  -H, --hmm_dir DIRECTORY  Path to a directory to store hmm-related data.
                           [required]
  -d, --download           If True, download the Pfam data from interpro.
  -q, --quiet              Disable verbose output.
  --path_pfam_a FILE       A path to downloaded Pfam-A HMM profiles. By
                           default, if `download` is ``False``,will try to
                           find it within the `hmm_dir`.
  --path_pfam_dat FILE     A path to downloaded Pfam-A (meta)data file. By
                           default, if `download` is ``False``,will try to
                           find it within the `hmm_dir`.
  -h, --help               Show this message and exit.

For the first-time usage, invoke

→ tkp-finder setup -H hmm -d

This will download Pfam-A HMMs and accompanying metadata, and split the models into categories. The resulting directory:

→ tree -L 2 hmm

hmm
├── PF00069.hmm
├── Pfam-A.hmm
├── Pfam-A.hmm.dat
├── pfam_entries.tsv
└── profiles
    ├── Coiled-coil
    ├── Disordered
    ├── Domain
    ├── Family
    ├── Motif
    ├── Repeat
    └── unknown

To dicover and annotate TKPs, refer to tkp-finder find command:

→ tkp-finder find --help

Usage: tkp-finder find [OPTIONS] [FASTA]...

Options:
  -H, --hmm_dir DIRECTORY    Directory with HMM profiles. Expected to contain
                             `profiles` dir and target PK profile
                             (PF00069.hmm). See `tkp-finder setup` on how to
                             prepare this dir.
  -t, --hmm_type TEXT        Which HMM types to use for annotating the
                             discovered TKPs. The names must correspond to
                             folders within he `hmm_dir`.  [default: Family,
                             Domain, Motif]
  -p, --pk_profile FILE      A path to the PK HMM profile. By default, will
                             try to find it within the `hmm_dir`.
  -m, --motif TEXT           A motif to discriminate between PKs and pseudo
                             PKs. This corresponds to the following conserved
                             elements::  (1) b3-Lys(2) aC-helix Glu(3-4-5) HRD
                             motif(6-7-8) DFG motif  [default: KEXXDDXX]
  -o, --output DIRECTORY     Output directory to store the results. Be
                             default, will store within `./tkp-finder`.
  -n, --num_proc INTEGER     The number of cpus for data parallelism: each
                             input fasta will be annotated within separate
                             process. HINT: one may split large fasta files
                             for faster processing.
  -q, --quiet                Disable logging and progress bar
  --pk_map_name TEXT         Use this name for the protein kinase domain.
                             [default: PK]
  --ppk_map_name TEXT        Use this name for pseudo protein kinases.
                             [default: PPK]
  --min_domain_size INTEGER  The minimum number of amino acid residues within
                             a PK domain.  [default: 150]
  --min_domains INTEGER      The number of domains to classify a protein as
                             TKP.
  --timeout INTEGER          For parallel processing, indicate timeout for
                             getting results of a single process.
  -h, --help                 Show this message and exit.