Zhang-et-al-project

Mapping meQTL-epi in ddm1-derived epiRILs in four-step QTL analysis

System requirements

• R version 4.0.2 with the following installed packages: data.table, qtl, dplyr, xlsx, stringr, ggplot2, reshape2

• Sufficient amount of RAM memory to run the QTL analysis (recommended minimum is 8 GB)

• The administrative privileges are required to install and run R‑Studio utilities.

• A network connection for data recovering over network.

Installation

To install the required packages type the following command in the R console.

install.packages(c('data.table', 'qtl', 'dplyr', 'xlsx', 'stringr', 'ggplot2', 'reshape2'))

To load the provided functions use load command in the R console.

source('functions/qtl-functions.R')

Demo

examplary_run.R file presents the demo of running the pipeline for toy datasets located in toy_dataset directory (containing phenotype.csv, genotype.csv, positions.csv and marker.csv files).

To run the demo file, use the R (RStudio is recommended), change the first row of the script by typing your working directory where you have placed all the files from this repository, eg:

setwd('/home/robert/Zhang-et-al-project')

and run the script in R console.

Expected output

The expected output is located in toy_outputs and it consists of the following files:

• PERM-all-traits_log_nm.Rdata - Rbinary dataset with the results from permutation test performed by rqtl package;

• QTL-direction-data-all-traits_log_nm.Rdata - Rbinary dataset with the effect direction for a given association (epimarker and phenotype);

• QTL-mapping-all-traits_log_nm.Rdata - Rbinary dataset with the results from the mapping procedure performed by rqtl package;

• REF-data-all-traits_log_nm.Rdata - Rbinary dataset with the crossing file used as the input for mapping procedure by rqtl package;

• SCANPOSINFO-all-traits_log_nm.Rdata - Rbinary dataset with positions of the markers after mapping procedure;

• gc_genotype_log_nm.csv - comma-separated csv file with the genotype data after preprocessing used for running the mapping procedure by rqtl package;

• mp_traits_log_nm.csv - comma-separated csv file with the phenotype data after preprocessing used for running the mapping procedure by rqtl package;

• info_peaks_log_nm.csv - comma-separated csv file with the positions of the peaks from mapping outputs.

Expected runtime

For 20 phenotypic traits, the whole pipeline takes around 1 minute, whereas:

• Running a mapping_qtl function could take 2 seconds per one trait (40 seconds for toy datasets);

• Running a getQTLpeaks_new function could take less than 1 second per one trait (10 seconds for toy datasets);

• Running a qtlpeaks.plot and transcis.plot functions could take no longer than 10 seconds in total.

Instructions for use

• Required input datasets:

phenotype dataset - csv comma divided file with quantitative traits for each line and phenotype

genotype dataset - csv comma divided file with epigenotype profile (M/U) for each epimarker and line

marker information dataset - directory to csv comma divided file with marker chromosome and start-end positions

positions_dir - directory to csv comma divided file with phenotype (epiRIL positions)

• Useful commands:

After collecting your datasets (make sure they have the same column names as in the toy datasets), load the required functions by using the following command:

library(data.table)
library(qtl)
library(dplyr)
library(xlsx)
library(stringr)
library(ggplot2)
library(reshape2)

and load the functions provided to this repository using:

source('functions/qtl-functions.R')

as the inputs you need to select your input/output directory and write down directories to your data, like below:

mp_traits_dir <- 'toy_dataset//phenotype.csv'
gc_genotype_dir <- 'toy_dataset//genotype.csv'
marker_dir  <- 'toy_dataset//markers.csv'
positions_dir <- 'toy_dataset//positions.csv'
output.dir <- 'examplary_outputs//'
input.dir <- output.dir 

Then you can use the functions to run the pipeline:

mapping_qtl(mp_traits_dir, gc_genotype_dir, marker_dir, output.dir)  
getQTLpeaks_new(input.dir, marker_dir)
plot1 <- qtlpeaks.plot(input.dir, marker_dir)
plot2 <- transcis.plot(input.dir, marker_dir, positions_dir)

By calling plot1 or plot2 in R with installed ggplot2 you can see your plots.

Functions overview

mapping_qtl function (required datasets: mp_traits_dir, gc_genotype_dir, marker_dir)

mapping_qtl - perform preprocessing of the data (based on the log transformation and preparing datasets for QTL analysis) - perform a mappinh and permutations (QTL analysis) - calculate genotype probabilites and map - calculate effect direction (phenotype ~ epigenotype)

getQTLpeaks_new function (required datasets: directory with outputs after performing mapping_qtl function)

getQTLpeaks_new - select a significant peaks with a given alpha level

qtlpeaks.plot function (required datasets: directory with outputs after performing mapping_qtl function, marker_dir)

qtlpeaks.plot - plot a LOD score thresholds for the peaks selected in getQTLpeaks_new

transcis.plot function (required datasets: directory with outputs after performing mapping_qtl function, marker_dir, positions_dir)

transcis.plot - plot a LOD score thresholds for the peaks selected in getQTLpeaks_new