fake.f2expr: Simulated expression traits for F2 intercross

In atbroman/neqtl: Tools for analyzing expression QTL by position on genome

Description

Simulated expression traits for an F2 intercross, obtained using sim.cross, with map10 as the map object.

Usage

data(fake.f2expr)

Format

An object of class cross. Seeread.cross for details.

Details

There are 500 F2 individuals with 1000 simulated expression traits. A random 10% are considered proximal (or CIS) traits, and a random 20% are considered distal or (or trans) traits.

See Also

sim.cross map10

Examples

## Not run: 
library(qtl)

# random seed
set.seed(5297830)

# load map
data(map10)

# drop X chr
map10 <- map10[-20]

# chromosome lengths
L <- summary(map10)[1:19,"length"]
names(L) <- as.character(1:19)

# simulate genes in different positions
n.genes <- 1000
chr <- factor(sample(names(L), n.genes, prob=L/sum(L), repl=TRUE), levels=1:19)
pos <- runif(n.genes, 0, L[chr])
genenam <- paste("gene", 1:n.genes, sep="")
genes <- data.frame(chr=chr, pos=pos)
rownames(genes) <- genenam


# which ones have cis?
cis <- sample(c(FALSE,TRUE), n.genes, prob=c(9,1), repl=TRUE)

# trans eqtl on chr 1 at 50 cM
trans <- sample(c(FALSE,TRUE), n.genes, prob=c(8,2), repl=TRUE)
trans[cis] <- FALSE # make cis ones be not trans

# cis effects
cis.eff <- runif(n.genes, 0.2, 4.9)
cis.eff[!cis] <- 0

# trans effects
trans.eff <- runif(n.genes, 0.2, 0.9)
trans.eff[!trans] <- 0

# simulate annotation
fake.annot <- data.frame(chr=factor(sample(names(L),
           n.genes, prob=L/sum(L), repl=TRUE), levels=1:19),
           pos=runif(n.genes, 0, L[chr]),row.names=genenam)
fake.annot[cis,"chr"] <- genes[cis,"chr"]
fake.annot[cis,"pos"] <-
  apply(cbind(genes[cis,"pos"]+runif(sum(cis),min=-2.5,max=2.5),
        L[genes[cis,"chr"]]),1,function(x) max(0,min(x)))

#save(fake.annot,file="neqtl/data/fake.annot.RData")

# simulate cross; qtl for all cis locations plus the one trans QTL
n.ind <- 500
x <- sim.cross(map10,
               model=rbind(c(1, 50, 0, 0), cbind(as.numeric(chr), pos, 0, 0)),
               type="f2", n.ind=n.ind)
origqtlchr <- c(1, chr)
origqtlpos <- c(50, pos)
o <- order(origqtlchr, origqtlpos)
# qtl get sorted by position so we have to go back
qtlgeno <- x$qtlgeno
qtlgeno[,o] <- qtlgeno # restore to original order
x <- clean(x) # drop qtl geno from object

# simulate phenotypes
pheno <- matrix(rnorm(n.ind*n.genes), ncol=n.genes)

# make them correlated
pheno <- pheno + rnorm(n.ind)

# add effects for cis traits
for(i in which(cis)) 
  pheno[,i] <- pheno[,i] + (qtlgeno[,i+1]-2)*cis.eff[i]

# add effects for trans traits
for(i in which(trans))
  pheno[,i] <- pheno[,i] + (qtlgeno[,1]-2)*trans.eff[i]

# add column names; make data frame; add to cross
colnames(pheno) <- genenam
x$pheno <- as.data.frame(pheno)

fake.f2expr <- x
#save(fake.f2expr,file="neqtl/data/fake.f2expr.RData")

## End(Not run)

data(fake.f2expr)
fake.f2expr <- calc.genoprob(fake.f2expr)
out <- scanone(fake.f2expr, method="hk", phe=1:nphe(fake.f2expr))

atbroman/neqtl documentation built on May 10, 2019, 2:08 p.m.