RNA Secondary Structure Prediction

Using dataset of Non Coding SNPs predict secondary structure with a variation of Nussinovs algorithm, energy_min.py RN_Analyze.py is run first and will use energy_min.py to return a .tsv file SEQ_DB.tsv with sequences and their corresponding dot bracket structures. Base pair distance is calculated using bp_distance from the RNA package taken from ViennaRNA. Visuals are created with RNAvisual.py.

To Run

First run: python RN_Analyze.py SNP.tsv sequences.fa

To run visualization: python RNAvisual.py SEQ_DB.tsv

Installations Required

RN_Analyze

Reading fasta file: pip install biopython Bp distance: miniconda or anaconda can be used

1. conda create -n viennarna -c bioconda viennarna
2. conda activate viennarna

Alternatively follow steps provided by viennarna or use homebrew: https://www.tbi.univie.ac.at/RNA/ViennaRNA/doc/html/install.html

RNAvisual

pip install dash pip install dash_bio pip install pandas

Dataset

Filtering all clinically significant SNPs

Biotypes

• mitochondrial tRNA
• micro RNA
• Small nucleolar RNA

Clinical significance

• Benign, likely benign, uncertain, likely pathogenic, pathogenic

SNP location and substitution

Compare remaining SNPs to all ncRNA to select appropriate genes

energy_min.py

Follows nussinovs algorithm with the following additions

1. Minimum loop

   This means that pairs must have some distance between them Ex. min loop of 2 would reject ..(.).. And accept .(..)..

2. Stacked base pairs

   ((.......)) preferred over (.(...)...)

3. Score Pairing

   GC preferred over AU and GU

RN_Analyze

run this first with

python RN\_Analyze SNP.tsv sequences.fa

outputs SEQ\_DB.tsv

RN_Analyze uses energy min to create dot structures for RNA sequences. This program takes 1-2 minutes to run

RNAVisual

Visual representations using dash to run this you must first run python RN\_Analyze.py SNP.tsv sequences.fa

after this run python RNAvisual.py SEQ\_DB.tsv

dash will run on http://127.0.0.1:8050/