Variational inference accelerates accurate DNA mixture deconvolution

Mateusz Susik

doi:10.1016/j.fsigen.2023.102890

Variational inference accelerates accurate DNA mixture deconvolution

2023-07-01

0

OA

AI

DOI：10.1016/j.fsigen.2023.102890

Original

Original request for help

Save

Abstract

En 中文

We investigate a class of DNA mixture deconvolution algorithms based on variational inference, and we show that this can significantly reduce computational runtimes with little or no effect on the accuracy and precision of the result. In particular, we consider Stein Variational Gradient Descent (SVGD) and Variational Inference (VI) with an evidence lower-bound objective. Both provide alternatives to the commonly used Markov-Chain Monte-Carlo methods for estimating the model posterior in Bayesian probabilistic genotyping. We demonstrate that both SVGD and VI significantly reduce computational costs over the current state of the art. Importantly, VI does so without sacrificing precision or accuracy, presenting an overall improvement over previously published methods.

Keywords:

Probabilistic genotyping

Stein variational gradient descent

Variational inference

Bayesian inference

Runtime

Precision

Variational inference accelerates accurate DNA mixture deconvolution

Abstract

Journal

Organization