Analysis Description
--------------------
Perform a Fast Unbiased AppRoximate Bayesian (FUBAR) analysis of a
coding sequence alignment to determine whether some sites have been
subject to pervasive purifying or diversifying selection. v2.1
introduces two more methods for estimating the posterior distribution of
grid weights: collapsed Gibbs MCMC (faster) and 0-th order Variation
Bayes approximation (fastest). Please note that a FUBAR analysis
generates a cache and a results JSON file in the same directory as
directory as the original alignment. HyPhy needs to have write
privileges to this directory. For example if the original file is in
/home/sergei/FUBAR/data/pol.nex then at the end of a FUBAR run, there
will also exist FUBAR-generated files
/home/sergei/FUBAR/data/pol.nex.FUBAR.json,
/home/sergei/FUBAR/data/pol.nex.fubrar.cache. They also provide
checkpointing so that a partially completed analysis can be restarted.

- __Requirements__: in-frame codon alignment (possibly partitioned) and a phylogenetic tree
(one per partition)

- __Citation__: FUBAR: a fast, unconstrained bayesian approximation for inferring
selection (2013), Mol Biol Evol. 30(5):1196-205

- __Written by__: Sergei L Kosakovsky Pond

- __Contact Information__: spond@temple.edu

- __Analysis Version__: 2.2


>code –> Universal

>Loaded a multiple sequence alignment with **119** sequences, **835** codons, and **1** partitions from `/home/datamonkey/datamonkey-js-server/production/app/fubar/output/5e85b99d89e95e52b24a986d`
Save FUBAR cache to : 
>cache –> /home/datamonkey/datamonkey-js-server/production/app/fubar/output/5e85b99d89e95e52b24a986d.FUBAR.cache
> FUBAR will write cache and result files to _/home/datamonkey/datamonkey-js-server/production/app/fubar/output/5e85b99d89e95e52b24a986d.FUBAR.cache_ and _/home/datamonkey/datamonkey-js-server/production/app/fubar/output/5e85b99d89e95e52b24a986d.FUBAR.json_, respectively 


> Number of grid points per dimension (total number is D^2) (permissible range = [5,50], default value = 20, integer): 
>output –> /home/datamonkey/datamonkey-js-server/production/app/fubar/output/5e85b99d89e95e52b24a986d.fubar
> Number of grid points per dimension (total number is D^2) (permissible range = [5,50], default value = 20, integer): 
>grid –> 20

>method –> Variational-Bayes
> The concentration parameter of the Dirichlet prior (permissible range = [0.001,1], default value = 0.5): 
>concentration_parameter –> 0.5


### Obtaining branch lengths and nucleotide substitution biases under the nucleotide GTR model
* Log(L) = -9075.31, AIC-c = 18637.01 (243 estimated parameters)
* Tree length (expected substitutions/site) for partition 1 :    0.365

### Computing the phylogenetic likelihood function on the grid 
* Determining appropriate tree scaling based on the best score from a  20 x 20 rate grid
* Best scaling achieved for 
	* synonymous rate =  2.815
	* non-synonymous rate =  0.357
* Computing conditional site likelihoods on a 20 x 20 rate grid