Add MSA workflow

[pjotrp]

@ekg is working on an MSA workflow

[pjotrp]

MSA is quadratic slow for growing datasets. So we need a way to split that data into smaller families. @ekg writes: You can script it out with mash and minhash distances from the pangenome.

[pjotrp]

@mr-c added a workflow at https://github.com/common-workflow-lab/2020-covid-19-bh/tree/master/msa. We need to split it up in subsections though - running MSA on all data is too slow.

[ekg]

Did you manage to get maffer to work?

…

On Sat, Apr 25, 2020, 14:58 Pjotr Prins ***@***.***> wrote: @mr-c <https://github.com/mr-c> added a workflow at https://github.com/common-workflow-lab/2020-covid-19-bh/tree/master/msa. We need to split it up in subsections though - running MSA on all data is too slow. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#11 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AABDQEORWDG7ZACIOC2DSXLROLM6BANCNFSM4MHCEGGA> .

[pjotrp]

We have not tried that yet on the GFA. Can you try and see if it works on
https://workbench.lugli.arvadosapi.com/collections/lugli-4zz18-z513nlpqm03hpca

[mr-c]

https://workbench.lugli.arvadosapi.com/container_requests/lugli-xvhdp-a2w5l0y4sh6efd1 is the latest run of mafft+iqtree, but iqtree didn't converge

[mr-c]

Running again with more maximum iterations https://workbench.lugli.arvadosapi.com/container_requests/lugli-xvhdp-xfh79fymt8v71d9#Log

[pjotrp]

@ekg we may need a maffer example to split the population into smaller subsets to run MSA on.

[pjotrp]

Maybe never.

[AndreaGuarracino]

maffer is currently deprecated.

@pjotrp, if we still want to have an MSA, in the current state of PubSeq, the fastest way to get it is to take advantage of the fact that we are currently apply SPOA on the whole input sequences. This can already give us a huge MSA, from which I can create a huge MAF output too.