QTL Effects after running scanone()

[venkat rami reddy]

Hi Karl

I have done the quick scanone to check the qtls in DH maize populations 

plot(scanone(cross,pheno.col = 14))

# Perform the permutation test
perm_results <- plot(scanone(cross, method="hk", pheno.col=14, n.perm=500))

# Summary of permutation results
summary(perm_results)

# Extract LOD scores from permutation results
perm_lod_scores <- perm_results$breaks

# Plot a single QTL scan with the permuted LOD scores
plot(scanone(cross, pheno.col = 14), col = "blue", main = "QTL Scan with Permutation Test", perm = perm_results)

# Add a red line for the permutation LOD threshold
abline(h = max(perm_results$breaks), col = "red", lty = 7)

# Set the LOD threshold based on permutation results
threshold <- max(perm_results$breaks)

# Perform a single QTL scan
qtl_results <- scanone(cross, pheno.col = 14)

# Identify significant QTLs based on the threshold
significant_qtls <- summary(qtl_results, threshold = threshold)

> print(significant_qtls)
               chr   pos  lod
chr3_207470669   3 59.40 9.30
chr7_4795874     7  8.61 5.84

How can i have the QTL effects for the significant SNPs 

thanks in advance 

Rami

[Karl Broman]

Use makeqtl() followed by fitqtl(), the latter with get.ests=TRUE.

You can get estimates for each in single-QTL models, or look at the estimates in a two-QTL, additive model.

qtl <- makeqtl(cross, c(3,7), c(59.40, 8.61), what="prob")

out_chr3 <- fitqtl(cross, pheno.col=14, qtl=qtl, formula=y~q1, get.ests=TRUE, method="hk")

out_chr7 <- fitqtl(cross, pheno.col=14, qtl=qtl, formula=y~q2, get.ests=TRUE, method="hk")

out_both <- fitqtl(cross, pheno.col=14, qtl=qtl, formula=y~q1+q2, get.ests=TRUE, method="hk")

summary(out_chr3)

summary(out_chr7)

summary(out_both)

karl

[Othmane Lamoumni]

Hello everyone!

I would just like to follow up with an encountered error when running fitqtl(get.ests = TRUE):

Error in solve.default(t(Z) %*% Z, t(Z) %*% X) : 

system is computationally singular: reciprocal condition number = 2.06693e-18

This singularity error is encountered in all my other analysis. Can this be a case?

Thank you

[Karl Broman]

This means the model is over-specified and cannot be fit.

There are too many QTL in the model, or some of the QTL have genotypes that are perfectly correlated.

karl

[Othmane Lamoumni]

I guess I can understand how QTLs with perfectly correlated genotypes can be considered "redundant" parameters in the model. However, it seems that the error persists even with limited QTLs (3 or 4). What can be a solution in this case? Perhaps reducing marker data (since I'm working using high-density maps)?

Othmane

[Karl Broman]

I don’t have enough information to be able to answer these questions.

karl

On Feb 5, 2024, at 6:58 AM, Othmane Lamoumni <lamou...@gmail.com> wrote:



--
You received this message because you are subscribed to the Google Groups "R/qtl discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rqtl-disc+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/rqtl-disc/b74117a6-7949-44f4-bd4d-0c66fb56627cn%40googlegroups.com.