Organization Acknowledgements
Senior Personnel Postdocs/Grad Students All Personnel by Institution
Collaborations
Presentations Publications
Community Software Morphology Databases Sequence Databases ATol

CIPRES Portal

The CIPRES Portal is currently under development. The portal is built on top of the CIPRES software libraries. The CIPRES project focuses on the inference of large trees. All software developed directly by the project is free and platform independent, but we are still working on a pain-free installation procedure. The portal lets you explore software that has been adapted to the CIPRES system without having to install it on your own machine. Calculations are carried out on the CIPRES cluster, at the San Diego Supercomputer Center.

Currently Available Tools for Large Tree Inference

Maximum Likelihood GARLI
Rec-I-DCM3 (GARLI)
RAxML
Rec-I-DCM3 (RAxML)
Parsimony PAUP
Rec-I-DCM3 (PAUP)
 
Bayesian MrBayes (Coming Soon)  
Simultaneous estimation of
trees and alignments:
Poy (Coming Soon) SATe (Coming Soon)

Note that each of the tree inference methods can be used as a standalone or in combination with a meta-heursitic to extend the range of problems that can be analyzed. The CIPRES portal is currently using Rec-I-DCM3 , a divide-and-conquer technique as a higher level algorithm. For details refer to the discussion of Disk Covering Methods (pdf) and a specific discussion of Rec-I-DCM3 ( pdf). An implementation of the Nixon's ratchet (link to publisher) will be available soon.

Requirements:
Input: A matrix of DNA (All) or AA (PAUP,RAxML) sequences in NEXUS or PHYLIP format. A starting tree is not required, but you can improve upon an initial estimate of a tree. To do this, simply include this estimate in a Trees block in the NEXUS file. If there are multiple trees in the Trees block, the first one will be used.

Limitations:
At this moment, configuration of parameters by users is not supported, rather, default parameters are used, based on the recommendations of each program's author. Users can deploy up to 4 active jobs at any one time. Because the cluster is a shared resource, jobs that run over 72 hours are subject to being killed before they return results.

Output:
Output is returned to the user as a file in valid NEXUS format. A link to this file is displayed in the browser (HTML), and mailed to the user.
Results will be stored for 7 DAYS ONLY, then will be purged.

Report an Issue with the Portal Request a Feature