compbiolab-CLI 0.4.0

CLI tool to compare and search protein families

Table Of Contents

• Table of Contents
• Software Info
• Overview
• Installation
• Usage
  • Comparing protein sequences
    • Comparing Arguments
  • Searching
    • Searching Arguments
• Optional Flags
• Available Metrics
• Examples

Software Info:

Software Name: compbiolab-CLI
Latest Software Version: v0.4.0
PyPI: https://pypi.org/project/compbiolab-CLI/
Documentation: https://cfogel.github.io/compbiolab-CLI-docs/

Overview

Our goal for this project is to read a new input sequence and find the protein family which it belongs in by comparing it with existing sequences in the database. In addition to finding the closest family for a new sequence, our search application can also accept an autoencoded fingerprint and see which family is the best match using a variety of different metrics.

Our compare application allows different protein families to be compared directly using different distance functions. It accepts the names of families already in the database, but it can also accept files containing data from new latent spaces.

The program can be downloaded from PyPI (the Python Package Index), and it has a Command Line Interface.

Installation

1. Install Python 3.7 or above
2. Open a terminal. (On Windows, Windows Key + R → Type cmd → Enter)
3. Type pip install compbiolab-CLI

Usage

Comparing protein sequences

Find the distance between fingerprints of two protein families

compare [-h] <protein_family> <protein_family> [distance_options] [output_options]
             list names

Comparing Arguments

• protein_family

  Protein family's name. Provide an existing protein family's name or the file name of a new latent space. Files should contain 30 floats, each float in a separate line.

  • distance options

    Optional distance flags

  • output_options

    Optional output flags

• list names

  Show available protein family names

Searching

Find the closest family to a new protein sequence or family

search [-h] lat <latent space> [distance_options] [output_options]
            seq <sequence> [output_options]
            list names

Searching Arguments

• lat <filename> [distance_options] [output_options]

  Provide the file name of one or more new protein family latent spaces. The closest protein family to these new latent spaces will be shown.

  • distance_options

    Optional distance flags

  • output_options

    Optional output flags

• seq <filename> [output_options] (Requires 64-bit Python 3.7.x)

  Provide the name of one or more files containing a protein sequence to get the closest protein families for those sequences.

  • output_options

    Optional output flags

• list names

  Show available protein family names

Optional Flags

Output Options

• -out

  Output filename

• -of

  Output format, text or csv. Default: text

• -om

  Output mode, a[ppend] or w[rite]. Default: a

Distance Options

• -m

  Distance metric. Default: euclidean

• -p

  Scalar. The p-norm to apply for Minkowski, weighted and unweighted. Default: 2

Available metrics

euclidean (default), minkowski, cityblock, sqeuclidean, cosine, correlation, hamming, jaccard, chebyshev, canberra, braycurtis, yule, dice, kulsinski, rogerstanimoto, russellrao, sokalmichener, sokalsneath

Examples

You can find the Euclidean distance between two families ATKA_ATKC and CDSA_RSEP by running the command:

compare ATKA_ATKC CDSA_RSEP

If you want to find the Cityblock distance between ATKA_ATKC and a new latent space stored at second_new_latent_example.txt, you can run the command:

compare ATKA_ATKC second_new_latent_example.txt -m cityblock

If you want to find the cosine distance between two new latent spaces stored at first_new_latent_example.txt and second_new_latent_example.txt, you can run the command:

compare first_new_latent_example.txt second_new_latent_example.txt -m cosine

---

You can find the closest protein family to first_new_latent_example.txt in cosine distance by running the command:

search lat first_new_latent_example.txt -m cosine

You can find the closest family to a new protein sequence (for example new_sequence_example.txt) by running:

search seq new_sequence_example.txt