Error in optimizing with phylosig lambda

[lilyb3]

I would like to calculate the phylogenetic signal for 2001 traits in a for loop with 81 taxa. I had no problem calculating the phylogenetic signal using Blomberg's K, but the code for lambda produces an error.

• Traits is the trait file.
• se.dn is the standard error in traits measured.
• tree_pruned is the phylogenetic tree with 81 traits.

Here is what I've done for K: for(i in 1:ncol(traits)){ trialtrait<-setNames(traits[,i], rownames(traits)) st_error <- setNames(se.dn[,i], rownames(se.dn)) phylosig_mean[[i]] <- phylosig(tree_pruned,trialtrait, test=TRUE,nsim=10000, se=st_error) trialtrait <- NULL st_error <-NULL }

Here is what I have tried to do for lambda:

• meansig2 and meanKval are from the K calculations above.

for(i in 1:ncol(traits)){ trialtrait<-setNames(traits[,i], rownames(traits)) st_error <- setNames(se.mean.dn1[,i], rownames(se.mean.dn1)) phylosiglambda_mean[[i]] <- phylosig(tree_pruned,trialtrait, method="lambda", test=TRUE, se=st_error, start= c(meansig2[i], meanKval[i]))

trialtrait <- NULL st_error <-NULL }

I've gotten the following errors:

• "In control$fnscale <- -1 : Coercing LHS to a list": I think I solved this one by adding "control$fnscale <- -1 " to control parameters
• Convergence=52
• ERROR: ABNORMAL_TERMINATION_IN_LNSRCH

So far, I've tried to change the following with lambda but nothing seems to work :

• Increasing maxit to 10000, -Playing with factr and pgtol per Chat GPT's suggestions. I went up to factr = 1e15 and pgtol = 1e-20 but this didn't work
• I've calculated lambda without standard error values or hypothesis testing, which doesn't produce errors. Then, I removed all traits that had lambda <.001 recalculated lambda with hypothesis testing and standard error. This didn't seem to help.
• I scaled the traits from -1 to 1 using the scale() function in R.

I've tried troubleshooting with ChatGPT quite a bit but nothing seems to work. Do you have any suggestions? Thank you!

[liamrevell]

Hi @lilyb3. I tried trouble-shooting this for you with simulated data & was able to get a for loop working without too much difficulty. Perhaps if you send or post your files I may be able to help more. -- Liam

[liamrevell]

Please re-open if you still need help with this issue.