Dear HyPhy creators,

I am using the BranchSiteREL.bf file to look for positive selection in some simulated and real dataset on a computer cluster.

The installation on the cluster worked well but had a little issue that did not appear when I made a first test-installation on my local machine (MacOSX 10.6.8). I use the inputRedirect method mentionned in the tutorial and it all works fine on my MacOSX, I just need to specify the full paths of all my files (tree, sequence, output and the BranchSiteREL.bf file itself), but it appears to find automatically the path to the loaded libraries ('LoadFunctionLibrary("chooseGeneticCode");' for example, which file I found in /usr/local/lib/hyphy/TemplateBatchFiles/TemplateModels/).

However going to the cluster and using the same inputRedirect method, I get the following error:

[code]
Error:Could not read command file in ExecuteAFile.
Original path: 'chooseGeneticCode'.
Expanded path:
'/stn4/ul/monthly/blaurenc/HyPhy/scripts//chooseGeneticCode'[/code]
I was wondering how I could specify to HYPHYMP the path to its own libraries, which where automatically found on my MacOSX ?

I also have another question : I found int the output of the BranchSiteREL.bf the information about which branch are under positive selection (in the console output and in the .out file) but I was wondering where I could find the information about which site is under positive selection and with what probability ? Is that information available ? Should I run something else to get it ?

Thanks a lot in advance for the answer :-)

Hi there,

I would hazard a guess that your cluster issue is due to the fact that HyPhy (by default) expects to find its library files in /usr/local/lib/hyphy
Please check to see where the "TemplateBatchFiles" directory lives on the cluster, or use the build-time options (README Multimedia File Viewing and Clickable Links are available for Registered Members only!!  You need to Login) to have HyPhy install somewhere else. Also, if you are using a queueing system (e.g. SGE or torque), it may only copy the HyPhy binary to the remote (compute) nodes. In this case, you may need to modify your job submission files to also "bring along" all the libraries that HyPhy has.

In terms of finding sites under episodic selection, we have an upcoming paper in PLoS Genetics (preprint attached), which describes the MEME method for doing this. The paper contains is a pertinent quote that I would like to draw your attention to:


Quote:

The ability of MEME, or similar substitution model-based methods, to accurately infer the identity of individual branches subject to diversifying selection at a given site seems unavoidably limited. Most of the information that such inference might be based on is limited to character substitutions along a single branch at a single site, i.e. one realization of the Markov substitution process. Selection along terminal branches in the context of negatively selected background can be detected more reliably than selection along interior branches among neutrally evolving background lineages. However, we caution that despite obvious interest in identifying specific branch-site combinations subject to diversifying selection, such inference is based on very limited data (the evolution of one codon along one branch), and cannot be recommended for purposes other than data exploration and result visualization. This observation could be codified as the “selection inference uncertainty principle” – one cannot simultaneously infer both the site and the branch subject to diversifying selection. In this manuscript, we describe how to infer the location of sites, pooling information over branches; previously [20] we have outlined a complementary approach to find selected branches by pooling information over sites.

Sergei

---

Dear Serguei,

First, thanks a lot for your reply.
We finally managed to install HyPhy on our cluster and it works fine.
Still, I would have one more little question about this, concerning the multi-threading feature of HyPhy. We compiled HyPhy with the MP2 target but for some benchmarking tests of HyPhy I would need it to use only a single thread. Is there a way to make HyPhy only use one core/CPU even if there are more available ?

Concerning the sites, your answer was helpful and your quote indeed really pertinent, thank you.
Still, if one would try to know which sites are *potentially* the ones under positive selection, may I ask you if there is any implementation of your MEME method yet in HyPhy or planned in an upcoming release ?

A last question : do you have an option for HyPhy to suppress computation on leaf branches ?

Thanks again a lot,


Cheers,

B. Laurenczy