weights/proportion of genes

[Lindsey Worcester]

Hi Claudia and Cècile,

I am running SNaQ on a RAD-seq. dataset, and I've been getting hybrid edges to show up, but not the weights/proportion of genes. I'm wondering what the issue could be. At first the weight did show up, but in all subsequent analyses, weights have been absent and on different analyses. I am using SNPs2CF for concordance factors, and feeding SNaQ a best tree from RAxML for the topology. 

This is what I've been doing (up to net 7 for one analysis:  

CF = readTableCF("SNPs2CF_mydataset.csv")

tre = readTopology("RAxML_bestTree.m10_20230712x6392SNaQtest1.tre")

net0 = snaq!(tre, CF, hmax=0, filename="/home/lindsey/Documents/Analyses/SNaQ/mydatasetmaxSNPsNULL/net0", seed=1)

net1 = snaq!(net0, CF, hmax=1, filename="/home/lindsey/Documents/Analyses/SNaQ/mydatasetmaxSNPsNULL/net1", seed=2)

Thanks!

Regards,

Lindsey

[Cécile Ané]

Hi Lindsey,

In some cases, the inheritance weights are not identifiable. In these cases, SNaQ uses composite parameters that are identifiable, but cannot estimate the inheritance weights separately from other parameters (like coalescent units of that branch or adjacent branches), and therefore does not return them. Some of your estimated networks might fall into that category.

I am puzzled by the fact that inheritance weights did show up at first, though. Perhaps on networks with different topologies?

Cécile.