Rule-Based Generation of Synthetic Genetic Circuits

EasyChair Preprint 9031

Abstract

Similar to the expandability of natural biological systems, that of synthetic biological systems is derived from the huge combinatorial search space of biological components, such as protein coding sequences and regulatory sequences. Due to this huge space, adequate design strategies are required for the implementation of synthetic genetic circuits in cells.

One design strategy for genetic circuits is a combination of sub-circuits. Recent progress in automated computational design has achieved multi-layered logic gates. Another direction of computational design can be reliance on expert knowledge. Indeed, even manual combinations of sub-circuits have allowed the implementation of prescribed cellular behavior.

To develop a support tool for genetic circuit design by biologists, here, we sought to combine inference machine and deep learning to generate and screen candidates of synthetic genetic circuits, respectively. Once an adequate rulebase is prepared, a logic programming language such as Prolog allows the designed cellular behavior to be broken down into combinations of rules, each understandable by a biologist. Simultaneously, each combination can indicate a genetic network topology from which published tools can estimate adequate parameters and suggest genetic parts. Although inference engines can potentially cause combinatorial explosions, using machine learning for candidate screening before the numerical simulation can circumvent this problem.

Keyphrases: Prolog, Rulebase, rule base, synthetic genetic circuit

@booklet{EasyChair:9031,
  author    = {Daisuke Kiga and Kazuteru Miyazaki and Shoya Yasuda and Ritsuki Hamada and Sota Okuda and Ryoji Sekine and Naoki Kodama and Masayuki Yamamura},
  title     = {Rule-Based Generation of Synthetic Genetic Circuits},
  howpublished = {EasyChair Preprint 9031},
  year      = {EasyChair, 2022}}