LD-proxy.Rmd

---
title: "LD-proxy"
author: "Gerardo José Rodríguez"
date: "2023-08-07"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(snpStats)
```


## Set your input variables
```{r}
# Determine what are the SNPs that you are interested to search in the datasets
SNPs.ref = c('rs3131962', 'rs2927718', 'rs183027878', 'rs76867316', 'rs3763196',
               'rs3858666', 'rs17158406', 'rs74758431', 'rs55957521', 'rs16882975')
  
# Define the dataset from where you want to obtain it
input_data = vroom('data/initial_data/Systolic blood pressure_GCST90000062.txt.gz')
```


## Read the PLINK files with genome info
```{r}
geno_panel = snpStats::read.plink(bed = 'data/genotype_reference/1kg.v3/EUR.bed', 
                                  bim = 'data/genotype_reference/1kg.v3/EUR.bim', 
                                  fam = 'data/genotype_reference/1kg.v3/EUR.fam')

```

## Make a clever object from your list of desired input SNPs to study

```{r}
# Return all sorted LD correlation measures of a SNP in a reference panel
build_LD_proxy_list = function(SNPref, plink_panel, window_kb=500, LD_measure='R'){

# Get the genomic location
gloc = geno_panel$map %>% filter(snp.name==SNPref) %>%
  dplyr::select(chromosome, position) %>% as.vector()

# Find closest SNPs (exclude reference)
SNPs = plink_panel$map %>% 
  filter(chromosome==gloc$chromosome) %>%
  filter(abs(position-gloc$position)<(window_kb*1000/2)) %>%
  filter(!snp.name==SNPref) %>%
  pull(snp.name)

# Identify their correlations, sort them (in absolute value) and return the vector
ldcorr = snpStats::ld(plink_panel$genotypes[,SNPref], 
             plink_panel$genotypes[,SNPs], 
             stats=LD_measure) 
ldcorr = setNames(as.vector(ldcorr), colnames(ldcorr))
return(ldcorr[order(abs(ldcorr), decreasing = T)])
}

listLD_leadSNPs = lapply(SNPs.ref, 
                         function(snp){build_LD_proxy_list(snp, geno_panel, 
                                                           5000, LD_measure = 'R')})

```



## Perform a search of potential SNPs to substute with proxy SNPs

```{r}

```