Debatable choices in sparse model

vlanore / diffsel

A C++ program to detect convergent evolution using Bayesian phylogenetic codon models.

Other

6 stars 2 forks source link

Debatable choices in sparse model #1

Closed pveber closed 5 years ago

pveber commented 7 years ago

[x] fitness inverse rate parameter (m_a): uniformDirichlet
[x] fitness shape parameter: 1
[x] hyperparameters for probconv (pi_k), lambda_k and mu_k: log normal(0,1)
[x] Random::beta from Random::gamma
[x] fitness[aastate] + 1e-6;
[x] moveFitness: move one random AA at each site for given cond?

vlanore commented 7 years ago

Regarding the beta from gamma point, what we meant is that we used this relation: https://math.stackexchange.com/a/190695 in order to draw from a beta distribution using the existing gamma implementation (as beta was absent from Random.cpp).

vlanore commented 5 years ago

Closing issue (outdated).