Incorporating temporal information during feature engineering bolsters emulation of spatio-temporal emergence.

MOTIVATION: Emergent biological dynamics derive from the evolution of lower-level spatial and temporal processes. A long-standing challenge for scientists and engineers is identifying simple low-level rules that give rise to complex higher-level dynamics. High-resolution biological data acquisition enables this identification and has evolved at a rapid pace for both experimental and computational approaches. Simultaneously harnessing the resolution and managing the expense of emerging technologies-e.g. live cell imaging, scRNAseq, agent-based models-requires a deeper understanding of how spatial and temporal axes impact biological systems. Effective emulation is a promising solution to manage the expense of increasingly complex high-resolution computational models. In this research, we focus on the emulation of a tumor microenvironment agent-based model to examine the relationship between spatial and temporal environment features, and emergent tumor properties. RESULTS: Despite significant feature engineering, we find limited predictive capacity of tumor properties from initial system representations. However, incorporating temporal information derived from intermediate simulation states dramatically improves the predictive performance of machine learning models. We train a deep-learning emulator on intermediate simulation states and observe promising enhancements over emulators trained solely on initial conditions. Our results underscore the importance of incorporating temporal information in the evaluation of spatio-temporal emergent behavior. Nevertheless, the emulators exhibit inconsistent performance, suggesting that the underlying model characterizes unique cell populations dynamics that are not easily replaced. AVAILABILITY AND IMPLEMENTATION: All source codes for the agent-based model, emulation, and analyses are publicly available at the corresponding DOIs: 10.5281/zenodo.10622155, 10.5281/zenodo.10611675, 10.5281/zenodo.10621244, respectively.

The in silico lab: Improving academic code using lessons from biology.

"Good code" is often regarded as a nebulous, impractical ideal. Common best practices toward improving code quality can be inaccessible to those without a rigorous computer science or software engineering background, contributing to a gap between advancing scientific research and FAIR practices. We seek to equip researchers with the necessary background and context to tackle the challenge of improving code quality in computational biology research using analogies from biology to synthesize why certain best practices are critical for advancing computational research. Improving code quality requires active stewardship; we encourage researchers to deliberately adopt and share practices that ensure reusability, repeatability, and reproducibility.

Fighting fire with fire: deploying complexity in computational modeling to effectively characterize complex biological systems.

Computational modeling empowers systems biologists to interrogate and understand increasingly complex biological phenomena, and the growing suite of computational approach presents both opportunities and challenges. Choosing the right computational approaches to address a given question requires managing a model's complexity, balancing goals and limitations including interpretability, data resolution, and computational cost. Excess model complexity can diminish the utility for building understanding, while excess simplicity can render the model insufficient for addressing the questions of interest. Using systems immunology as a case study, we review how different model design strategies uniquely manage complexity, ending with a consideration of composite models, which combine the benefits of individual paradigms but present additional challenges arising from added layers of complexity. We anticipate that considering general model design challenges and potential solutions through the lens of complexity will foster enhanced collaboration among computational and experimental researchers.

Automated Transition State Theory Calculations for High-Throughput Kinetics.

A scarcity of known chemical kinetic parameters leads to the use of many reaction rate estimates, which are not always sufficiently accurate, in the construction of detailed kinetic models. To reduce the reliance on these estimates and improve the accuracy of predictive kinetic models, we have developed a high-throughput, fully automated, reaction rate calculation method, AutoTST. The algorithm integrates automated saddle-point geometry search methods and a canonical transition state theory kinetics calculator. The automatically calculated reaction rates compare favorably to existing estimated rates. Comparison against high level theoretical calculations show the new automated method performs better than rate estimates when the estimate is made by a poor analogy. The method will improve by accounting for internal rotor contributions and by improving methods to determine molecular symmetry.