The engineering of synthetic genomes—complete or partial genome constructs designed de novo or heavily modified—represents a frontier in synthetic biology with vast potential for biotechnology, medicine, agriculture and basic science. However, traditional synthetic genome design remains highly labour‑intensive, reliant on manual annotation, iterative assembly and trial‑and‑error testing of large sequence spaces. Artificial intelligence (AI) offers a transformative opportunity to augment and accelerate synthetic genome design by learning patterns of genomic architecture, regulatory logic and functional outcomes. In this paper, we explore the development of AI‑based methodologies for designing and engineering synthetic genomes. We propose a workflow integrating generative models (e.g., language‑model architectures trained on genome sequences), predictive deep learning for functional annotation, and optimisation algorithms for minimal genome design, modular genome construction and error‑correction. To illustrate potential benefit, we present a hypothetical benchmarking study in which an AI‑augmented design workflow reduced design cycles by ~40 %, increased predicted functional genomic modules by ~30 %, and lowered predicted assembly error rate by ~25 % compared to manual design. Tabulated results highlight comparisons across methods. We discuss methodological elements including feature encoding (e.g., genomic motifs, regulatory elements, three‑dimensional chromatin structure), model architectures (transformer‑based sequence models, graph neural networks for genome assembly graphs), training/validation workflows, and deployment considerations (interpretability, biosafety, ethical oversight). Key challenges include limited training data for synthetic genomes, ensuring viability of AI‑designed genomes in living systems, and governance of biosecurity risks. Looking forward, we delineate future perspectives: foundation‑models for synthetic life, closed‑loop design‑build‑test‑learn pipelines with AI, and integration of synthetic genomes into programmable cellular factories. In conclusion, AI‑driven design and engineering of synthetic genomes provides a promising paradigm to accelerate synthetic biology, but must be advanced with rigorous experimental validation, robust modelling and responsible stewardship
Sikkander et al. (Wed,) studied this question.