leezx / varianttogene

Genomic predictors and Annotation predictors

VariantToGene

Genomic predictors and Annotation predictors

Genomic predictors | variant to gene | 变异的基因注释

本仓库为以上教程的代码实现

general input: a dataframe, rowname is SNP rsid, rsid column is rsid

general output: add annotation to the columns one by one, like nGene, eGene and cGene

prepare curated database

• this is the most important part
• the functions are very simple

Genomic predictors

• get_nGene(), Nearby (nGene)
• get_eGene(), eQTL (eGene)
• get_cGene(), Conformation (cGene)

Annotation predictors

• get_fGene(), Function (fGene)
• get_pGene(), Phenotype (pGene)
• get_dGene(), Disease (dGene)

Other functions

• extend_LD(), get all high LD (R2>0.8) SNPs

get_nGene

input:

• rsid df, a dataframe, rowname is SNP rsid, rsid column is rsid

data preparation:

• go to VEP website: https://asia.ensembl.org/Tools/VEP
• get the VEP annotation result, example: VEP/all.reported.snp.LD.merged_VEP.txt

function pipeline:

• simplify columns
• rename categories from VEP
• remove duplicate according to priority

output:

• unique function annotation of each SNP - a dataframe
• nGene_type: Splicing
• nGene: MYSM1
• nGene_biotype: protein_coding

get_eGene

input:

• rsid df, a dataframe, rowname is SNP rsid, rsid column is rsid

data preparation:

• go to eQTLGen download eQTL data: https://www.eqtlgen.org/cis-eqtls.html

function pipeline:

• simply match rsid to multiple genes

output:

• eGene: geneA,geneB

get_cGene

input:

• rsid df, a dataframe, rowname is SNP rsid, rsid column is rsid

extend_LD

input:

• rsid, example: unique(asso.df$SNPS)

function pipeline:

• library("LDlinkR")
• LDproxy()

output:

• all high LD SNPs - a dataframe