SNPeBoT2

SNPeBoT2 is a downloadable package for predicting the effect of SNPs on transcription factor binding and disease association.

---

Table of Contents

• Installation
• Usage
  • Predict SNP Effect
  • Run Association Analysis

---

Installation

1. Download this repository

2. Install ModCRElib inside the SNPeBoT2 folder by following the instructions at: https://github.com/structuralbioinformatics/ModCRElib

   • After installation, a ModCRElib/ folder containing all necessary files should be present in the current directory
   • Ensure all ModCRElib dependencies are met

3. Download required data files into the AssociationFilter/ folder and decompress them:

   • TFLink_Homo_sapiens_interactions_All_simpleFormat_v1.0.tsv.gz
   • AllPredictions.AvgHiC.ABC0.015.minus150.ForABCPaperV3.txt.gz
   • disgenet/all_gene_disease_pmid_associations.tsv.gz Store in AssociationFilter/disgenet/, An account is required to access disgenet (Free for academic purposes)

---

Usage

Predict_SNP_effect.sh

Generates allele-specific binding (ASB) predictions for a list of reference and alternate sequences for a given transcription factor.

Arguments:

#	Argument	Description
1	TF name	Name or ID of the transcription factor (should match the ID used in arguments 3 and 4)
2	job_id	A unique job ID used to name output folders
3	ASB file path	Path to the file containing the SNPs to be tested (see format below)
4	TF model	Path to the model of the TF bound to DNA (can be generated with ModCRElib or another tool)
5	template ID	ID of the template used in the model. If ModCRElib was used, this can be found in the model filename. Otherwise, use general

ASB input file format (4 columns per SNP, tab-separated):

Column	Description
1	TF name
2	Reference sequence (51 nt, mutation at position 26)
3	Alternate sequence (51 nt, mutation at position 26)
4	SNP ID (recommended format: {chr}-{position}-{ref allele}-{alt allele}-{TF name})

Example row:

FIGLA   GGATTTATCCATGTTTTTGCATGTAGAGATGGCTTGAAAAACAAACTATAT   GGATTTATCCATGTTTTTGCATGTACAGATGGCTTGAAAAACAAACTATAT   chr7-133924858-G-C-FIGLA

Output:

• Models submitted by the user are stored in Output_{job_id}/models_{job_id}/
• SNP binding effect predictions are stored in Output_{job_id}/Predictions_{job_id}.tsv

---

RunAssoc.sh

Predicts whether a SNP is associated with a disease, based on the output of Predict_SNP_effect.sh.

Arguments:

#	Argument	Description
1	TF name	UniProt ID of the transcription factor
2	Chrom	Chromosome on which the SNP is located
3	Prediction	Predicted binding effect from Predict_SNP_effect.sh (gain, loss, or no-change)
4	Position	Genomic position of the SNP
5	job_id	The job ID used in Predict_SNP_effect.sh
6	Assembly	Human genome assembly used for the SNP location (e.g. hg38)
7	SNP ID	The SNP ID as used in Predict_SNP_effect.sh

Output:

• Results are stored in Output_{job_id}/associations/assoc_{SNP_ID}.csv