Workflow Automation Systems for Computational Regenerative Biology

Abstract

Computational regenerative biology research involves complex multi-step analytical workflows -- from raw sequencing data through quality control, alignment, quantification, integration, machine learning analysis, simulation, and visualisation -- that must be executed reliably, reproducibly, and efficiently across diverse computing environments. Manual execution of these workflows through command-line scripts and interactive notebooks creates reproducibility failures, undocumented parameter choices, and inefficient re-execution of unchanged analysis steps. Workflow automation systems address these issues through declarative workflow specification, dependency tracking, caching of intermediate results, and automated re-execution on input changes. This paper proposes the Computational Regenerative Biology Workflow Automation (CRBWA) framework, a comprehensive workflow automation system integrating the full regenerative biology computational ecosystem into automated, reproducible, and intelligent analysis pipelines, comprising five automation components: a declarative workflow specification language (DWSL) for regenerative biology analysis graphs; an intelligent dependency tracker and cache manager (IDTCM) minimising redundant computation through smart caching; a cross-platform execution adapter (CPEA) enabling the same workflow specification to run on laptop, HPC cluster, and cloud without modification; a workflow quality assurance monitor (WQAM) detecting analysis anomalies and parameter drift in automated pipelines; and an AI-assisted workflow composer (AWC) using LLM-based assistance to help researchers specify and debug complex analysis workflows. CRBWA is evaluated on 18 benchmark analysis workflows from the regenerative biology ecosystem covering scRNA-seq, multi- omics integration, pathway simulation, tissue image analysis, and digital twin personalisation. DWSL reduces workflow specification effort by 68.4% vs. script-based approaches. IDTCM achieves 84.6% cache hit rate, reducing total compute time by 64.2%. CPEA enables cross-platform portability with mean 2.4% performance overhead. AWC assists researchers in correctly specifying 94.6% of workflows on first attempt. The study contributes the CRBWA specification, 18 benchmark workflow templates, and empirical evidence for workflow automation substantially improving regenerative biology research efficiency and reproducibility.