v2.1.0

What's Changed

• new error model by @m-murphy in #2

Full Changelog: 2.0.0...2.1.0

v2.0.0

This release comes with a newly revamped model. Previously, we were using a pseudo-marginal Metropolis Hastings approach to sample from a fully marginalized latent genetic space. Unfortunately, this did not perform very well in the context of high diversity genetic loci. For this release, we take the approach of fully marginalizing loci up to a user defined complexity limit (complexity_limit in the call to run_mcmc). Beyond this value, we instead augment the model with a latent genetic state which is subsequently sampled during the MCMC algorithm. This has resulted in substantially more stable estimates of allele frequencies and complexity of infection.

v1.0.0

Initial release of the MCMC implementation