Kris Varun Parag

Estimating spatially adjusted temperature-dependent time-varying reproduction numbers for vector-borne diseases

This paper estimates spatially adjusted temperature-dependent reproduction numbers for vector-borne diseases using simulations and real dengue data, finding that spatial weighting and temperature dependence improve Rt estimation accuracy.

A computationally tractable birth-death model that combines phylogenetic and epidemiological data

The paper introduces a computationally efficient birth-death model that integrates phylogenetic and epidemiological data to estimate transmission dynamics and unreported infections.

Are skyline plot-based demographic estimates overly dependent on smoothing prior assumptions?

This paper evaluates how much smoothing priors affect demographic estimates in Bayesian phylogenetics and introduces a statistic to measure their impact.

Optimising Renewal Models for Real-Time Epidemic Prediction and Estimation

This paper optimizes the renewal model for real-time epidemic prediction by developing a principled method to select the optimal window length k using the accumulated prediction error metric.

Adaptive Estimation for Epidemic Renewal and Phylogenetic Skyline Models

This paper introduces a method to optimally select the number of parameters in epidemic renewal and phylogenetic skyline models using minimum description length to improve estimation accuracy.

Jointly inferring the dynamics of population size and sampling intensity from molecular sequences

This paper introduces a new method to jointly estimate population size and sampling intensity dynamics from molecular sequences, showing improved accuracy over existing methods.

On Signalling and Estimation Limits for Molecular Birth-Processes

The paper shows that pathway capacity is not a reliable measure of precision in molecular birth-process networks and introduces a better-performing signalling motif called birth-following.

Robust Design for Coalescent Model Inference

This paper develops a robust experimental design framework for coalescent model inference by proving that logarithmic parameter transformation and uniform distribution of coalescent events minimises estimation uncertainty.

Optimal Point Process Filtering for Birth-Death Model Estimation

This paper introduces the Bayesian Snyder filter for estimating birth-death process parameters from phylogenies, showing it is mean square error optimal and comparable to existing methods.