ProNA

A machine-learning based model for predicting interactions between bacterial sRNAs and proteins. ProNA takes two input fasta file: one with protein sequences and another with RNA sequences. RNA sequences must have only A, C, G, and T. The output is a csv file which the probabilities of interaction and not-interaction for all possible pairings from the two input fasta files (i.e., if there are two protein sequences and ten sRNAs sequences then the output file will have twenty rows).

Setup

For running ProNA, you can download Anaconda from here

Requirements

• Python 3.9.13
• os 3.9.13
• Collections 3.9.13
• Numpy 1.21.5
• Math 3.9.13
• Random 3.9.13
• Pandas 1.4.4
• Joblib 1.1.0
• xgboost 1.7.3

Usage

To run ProNA, you need a fasta file with the sRNA sequences and then a fasta file with the protein sequences. For example sRNA fasta file with two sequences might look like this:

>2IY5-T
GCCGAGGTAGCTCAGTTGGTAGAGCATGCGACTGAAAATCGCAGTGTCCGCGGTTCGATTCCGCGCCTCGGCACCA
>1TRN-R
AATCCATTGCACTCCGGATTT

and a protein fasta file with two sequences migth look like this:

>3CW1-O
MKLVRFLMKLSHETVTIELKNGTQVHGTITGVDVSMNTHLKAVKMTLKNREPVQLETLSIRGNNIRYFILPDSLPLDTLLVDVEPKVKSKKREAVAGRGRGRGRGRGRGRGRGRGGPRR
>3CW1-N
MKLVRFLMKLSHETVTIELKNGTQVHGTITGVDVSMNTHLKAVKMTLKNREPVQLETLSIRGNNIRYFILPDSLPLDTLLVDVEPKVKSKKREAVAGRGRGRGRGRGRGRGRGRGGPRR

In the command line, type:

python3 ProNA.py

Then you will be prompted to enter the fasta files:

Enter your sRNA file: sRNAs.fasta
Enter your protein file: Proteins.fasta

Result

ProNA result consists of:

• Rows (4 rows): sRNA ID, protein ID, Probability of interaction, and Probability of non-interaction.
• Columns: RNA sequences times protein sequences.

Note: The output is sorted based on probability of interaction. Here there are few sample rows from ProNA output:

2IY5-T  3CW1-O  0.9571418 0.0428581
2IY5-T  3CW1-N  0.6550656 0.3449343
1TRN-R  3CW1-O  0.4539958 0.5460042
1TRN-R  3CW1-N  0.1448176 0.8551823