Performing A Multiple Sequence Alignment And Ml Tree Construction With An Insufficient Amount Of Ram.

Question

[Caitlin]

Hi all.

I am attempting to perform a multiple sequence alignment followed by the construction of a maximum-likelihood tree but I am, currently, restricted to the use of an old machine with a insufficient amount of RAM (512MB). I have attempted to locate a web server on which to upload the 635 sequences but I have encountered two separate problems. Either my file exceeds the 1 MB limit or the fact that several of my sequences (in fasta format) have the word 'cadherin' in the identifier line, i.e., >cadherin...

Does anyone know of a online tool I could use? My local install of Muscle (3.8) and MEGA (5.2.2) met with the same memory issue as described above. My classes start in August so I could wait and see if the problem persists once I have access to the computer lab.

Thanks all.

Caitlin

Answer 1

[DG]

The CIPRES web-portal has some great access to a large computing cluster pre-installed with all sorts of phylogenetic inference software (not sure about the alignment step though, you'll have to check)

CIPRES

You may also really want to look in to whether there is any sort of cluster or supercomputing environment available at your institution. Email some people in your computer science, physics, and/or math/stats departments. You may also have some infrastructure more regionally that you can gain access too. For instance here in Canada we have the ComputeCanada regional network of supercomputing clusters that any academic at a University can gain access too.

There is also GALAXY, you should at the very least be able to do the alignment step there. They have some ML tools available but maybe not the exact program you want.

And finally there are all of the Amazon Instances you could pay for to use. Lots of pre-built bioinformatics ones out there that should have the resources you need.

[Manu Prestat]

I think also that Galaxy could handle both, didn't know about CIPRES.

Answer 2

[Pappu]

Use fasttree for this purpose.

[DG]

Pretty sure even FastTree may have issues with only 512 MB...