Use of CIPRES functionality

[rvosa]

In the review stage of the Syst. Biol. manuscript a remark was made to the effect that users without access to HPC resources should be able to use SUPERSMART in such a way that this takes advantage of the CIPRES hardware.

In order to explore this issue we have contacted Mark Miller, the PI of CIPRES at SDSC. He noted that there are basically two approaches for this:

1. install SUPERSMART on their hardware. However, this would give them (SDSC) the headache of having to manage SUPERSMART's resource consumption, e.g. queueing, drive space usage, etc., and we would have to test and develop further our abilities to interact with SUPERSMART remotely, e.g. using BioVeL/Taverna bindings.
2. we create the option of farming out individual steps, e.g. alignment or tree inference, on their servers. This means that users still have a local install of SUPERSMART, but some of the analyses are done remotely.

It appears that option 2 is preferable: less work for CIPRES as well as more granular local control for users. However, it would still be quite a bit of work to write client code to interact with CIPRES's REST interface, and to deal with their timeouts and such. We therefore will postpone this work for version 2.

[rvosa]

Note that there is a Bio::CIPRES client library on CPAN.