MiCFiG

MiCFiG (MitoCarta Filter GFF) is a bioinformatics tool that uses the MitoCarta genomic database to to filter GFF files of nonmodel organisms.

Contents

• Context
• Methods
  • Scripts
  • Pipeline
• Walkthrough
  • Installation
  • Requirements
  • Steps
  • Final Products

Context

A common practice in bioinformatics is to annotate genomic files. Annotation files (e.g. BED & GFF) contain the locations of known genes in genomic files, acting as a road map for genes of interest. By their definiton, the genomes of model organisms are incredibly well-annotated - lots of people have taken to the call and had their work double- & triple-checked. However, the pool of non-model organisms without fully annotated genomes is vast.

MiCFiG aims to make annotation of species with incomprehensive genetic labeling a much easier process. It does so by using BLAST to identify & filter genetic features in non-model organisms based on homologies with model species. The MitoCarta database, a catalog of mitochondrial & nuclear protein-encoding genes in Homo sapiens & Mus musculus, provides a diving board for such comparisons.

Methods

The MiCFiG pipeline is intended for non-model organisms with genomes that are not thoroughly annotated with informative gene names. By using the MitoCarta database, the seqeunces of known human genes are used as queries for non-model organisms to identify mitochondrial-targeting genes. These are then extracted in a filtered annotation file & named, providing annotated genes for downstream analysis.

Scripts

• split_fasta.py

  • Takes a comprehensive FASTA file (in this case, taken from the MitoCarta database) and splits it into individual FASTA files - one for each sequence

• blast_script.sh

  • Loops through FASTA files and performs a BLAST search

• process_blast_results.sh

  • For each BLAST hit, grabs the minimum subject start and maximum subject end from the top rated HSP score
  • process_blast.py
    • Processes BLAST results and makes a BED file for the top hits based on HSP score

• filter_gff.py

  • Identifies overlaps between BED & GFF files
  • Determines which gene from model organism belongs with the genes in non-model organism

• add_gene_name.sh ( I think this goes here in the pipeline )

  • Adds MitoCarta gene name to GFF file
    • Accounts for potential differences in genetic nomenclature between species
  • extract_gene_name.py
    • Extracts gene name & sequence ID from MitoCarta-produced FASTA file & puts it in a lovely text file

• process_gff.py

  • Creates complete & annotated GFF file for non-model organism
  • Keeps track of genes not found to have a sister homology

Pipeline

Walkthrough

Installation

To clone this repository using HTTPS, enter the following into the command line:

git clone https://github.com/KLab-UT/MiCFiG.git

To clone this repository using SSH, enter the following into the command line:

git clone git@github.com:KLab-UT/MiCFiG.git

Requirements

MiCFiG as it stands is computationally and time intensive. Using the scripts will require a high performance computer in order to function in a timely manner and without frying your machine.

MiCFiG will require a starting annotated genome as a GFF file, as well as a FASTA file containing the sequences of the genome. Everything else is provided by the repository.

WARNING: The scripts meant to run MiCFiG are currently specified for the test organism, Aspidoscelis marmoratus. We are still working on making the scripts generalized to run for any non-model organism.

Steps

Before running main.sh, you need to take the target organism's genomic sequences in FASTA format and turn them into a BLAST database. This needs BLAST installed on your device, which can be accomplished by running the following code:

makeblastdb -in your_sequences.fasta -dbtype nucl -out your_database_name

With your BLAST database, you can now run main.sh:

bash main.sh

Final Products

Running MiCFiG will create a working directory and the following:

• fastas/proteins

  • Found in working directory
  • Directory
  • Contains the sequences for each protein from MitoCarta in individual files

• blast_results/tblastn_output

  • Found in working directory
  • Directory
  • Results from tblastn

• tblastn_bed_files

  • Found in working directory
  • Directory
  • BED files generated for each BLAST result

• gene_ids.csv

  • Found in repository
  • File
  • Contains genes and their associated IDs

• tblastn_gff_filtered

  • Found in working directory
  • Directory
  • Contains the filtered GFFs of each gene

• tblastn_gff_filtered/not_found.txt

  • Found in working directory
  • File
  • Keeps track of genes not found when creating final processed GFF file

• tblastn_log.csv

  • Found in repository
  • File
  • Logs ID, gene, and annotation

• tblastn_merged.gff

  • Found in repository
  • File
  • Final annotated genome for non-model organism