Genes are subsequences within transcripts that can be translated into proteins by ribosomes. They contain a reading frame made up of consecutive triplets, starting with an initiation codon (e.g., 'AUG', 'UUG', 'CUG', 'AUU', or 'GUG') and ending with a stop codon (either 'UAA', 'UAG', or 'UGA'). All codons must be in the same reading frame.

Upstream of the initiation codon, there is typically a motif that aids in translation initiation by allowing the binding of the 16S ribosomal RNA subunit. This sequence, known as the Shine-Dalgarno sequence (AGGAGGUAA), helps position the ribosome correctly [Shine and Dalgarno, 1973]. It is worth noting that the Shine-Dalgarno motif doesn't necessarily have to be in the same reading frame as the initiation codon, and it may sometimes be incomplete.

While many organisms still lack experimentally verified genome annotations, gene prediction remains crucial for the automatic annotation of genomes. Several tools and methods are available for this task [List of gene prediction software].

In this project, we aim to develop a simple approach to predict prokaryotic genes by detecting reading frames and identifying the Shine-Dalgarno motif. Our objective is to predict genes for the reference genome of Listeria monocytogenes EGD-e (sequenced by the Institut Pasteur), which has 2,867 known genes. You can find the genome here.

To set up the environment and install the required dependencies, use the following commands:

conda env create -f environment.yml
conda activate gene-prediction

First, clone the repository and navigate to the project folder:

git clone git@github.com:zhukovanadezhda/gene-prediction.git
cd gene-prediction

To run the gene prediction script, use the following command:

python3 gpred/gpred.py -i data/listeria.fna \
                       -p results/predicted_gene_positions.csv \
                       -o results/predicted_genes.fasta

The following command will analyze the Listeria genome file located at data/listeria.fna to predict genes. It specifies a minimum gene length of 50 nucleotides, a maximum allowable distance of 16 nucleotides between the start codon and the Shine-Dalgarno sequence, and requires at least a 40-nucleotide gap between consecutive genes. The predicted gene sequences and their corresponding positions will be saved in results/predicted_genes_positions.csv and results/predicted_genes.fasta, respectively.

  -h, --help                      Show help message and exit
  -i GENOME_FILE                  Complete genome file in FASTA format
  -g MIN_GENE_LEN                 Minimum gene length to consider (default: 50)
  -s MAX_SHINE_DALGARNO_DISTANCE  Max distance from start codon to search for Shine-Dalgarno motif (default: 16)
  -d MIN_GAP                      Minimum gap between two genes (excluding Shine-Dalgarno box, default: 40)
  -p PREDICTED_GENES_FILE         Output CSV file with predicted gene positions
  -o FASTA_FILE                   Output FASTA file containing predicted gene sequences

To assess the accuracy of our gene predictions, we can compare them against a reference set of known genes using jvenn, an online tool developed by INRA for Venn diagram analysis. The comparison involves the following datasets:

1. Predicted genes: Genes identified by our program.
2. Reference genes (Prodigal): From data/prodigal.csv, which contains genes predicted by the Prodigal software.
3. Positions of reference genes: From data/positions.csv, detailing the location of known genes within the Listeria genome.

This comparison reveals that relying solely on the Shine-Dalgarno motif for gene prediction is insufficient to capture all genes accurately. However, it can still help identify certain genes that Prodigal does not predict.