Leonard Schmiester

Mini-batch optimization enables training of ODE models on large-scale datasets

This paper applies mini-batch optimization to ODE models, improving efficiency and results on a cancer signaling dataset.

COVID-19 Disease Map, a computational knowledge repository of SARS-CoV-2 virus-host interaction mechanisms

The paper describes the development of a computational knowledge repository, the COVID-19 Disease Map, which curates SARS-CoV-2 virus-host interaction mechanisms and provides tools for their analysis and application.

Cell-to-cell variability in JAK2/STAT5 pathway components and cytoplasmic volumes define survival threshold in erythroid progenitor cells

The paper models single-cell variability in JAK2/STAT5 signaling and cytoplasmic volume to explain how erythroid progenitor cells transition from binary survival/apoptosis decisions to graded population responses upon Epo stimulation.

Efficient gradient-based parameter estimation for dynamic models using qualitative data

This paper introduces a framework for efficiently estimating parameters of dynamic models from qualitative data using gradient-based optimization, achieving significant speedup and improved results.

Benchmarking of numerical integration methods for ODE models of biological systems

This study benchmarks numerical integration methods for ODE models in systems biology, finding that most ODEs are stiff and providing guidelines for algorithm and hyperparameter selection.

Statistical inference of mechanistic models from qualitative data using an efficient optimal scaling approach

The paper develops an efficient optimal scaling approach for parameter inference in dynamical models using qualitative data, improving robustness and reducing computation time.

Efficient parameterization of large-scale dynamic models based on relative measurements

The paper introduces a hierarchical method for efficiently parameterizing large-scale dynamic models using relative measurements, improving optimizer performance and enabling unbiased weighting of heterogeneous data.

Efficient parameterization of large-scale mechanistic models enables drug response prediction for cancer cell lines

The paper presents a computational framework to parameterize large-scale mechanistic models for predicting drug response in cancer cell lines using exome and transcriptome data, achieving higher accuracy than statistical approaches.