Deep learning from phylogenies to uncover the epidemiological dynamics of outbreaks.

Widely applicable, accurate and fast inference methods in phylodynamics are needed to fully profit from the richness of genetic data in uncovering the dynamics of epidemics. Standard methods, including maximum-likelihood and Bayesian approaches, generally rely on complex mathematical formulae and approximations, and do not scale with dataset size. We develop a likelihood-free, simulation-based approach, which combines deep learning with (1) a large set of summary statistics measured on phylogenies or (2) a complete and compact representation of trees, which avoids potential limitations of summary statistics and applies to any phylodynamics model. Our method enables both model selection and estimation of epidemiological parameters from very large phylogenies. We demonstrate its speed and accuracy on simulated data, where it performs better than the state-of-the-art methods. To illustrate its applicability, we assess the dynamics induced by superspreading individuals in an HIV dataset of men-having-sex-with-men in Zurich. Our tool PhyloDeep is available on github.com/evolbioinfo/phylodeep .

Mechanistic studies on the ethyl-esterification of acitretin by human liver preparations in vitro.

Studies have been performed with human liver microsome preparations in vitro, to investigate the reaction mechanisms involved in the conversion of acitretin to the corresponding ethyl ester, etretinate. The results indicate that: Three fresh samples of human liver, which had been stored in liquid nitrogen for up to 8 months, all produced traces of etretinate (5.8 +/- 0.8 ng/ml) in the presence of ethanol but not when the acitretin was added in acetone, or when the sample was denatured by preheating. Studies with pooled human liver microsomes, to identify the cellular location of the enzymes and the co-factors involved in this esterification, indicate a primary requirement for both ethanol and CoA + ATP with a secondary potentiation in the presence of an NADPH regenerating system. A possible explanation for these finding is that the microsomal ligase enzymes form an intermediate ester between CoA and acitretin, which is then trans-esterified by the ethanol. The low formation with CoA + ATP may indicate that second stage of this process occurs spontaneously, with the NADPH potentiation suggesting that it could also be mediated enzymically.