Computer aided engineering of biological systems

Abstract

The rapid growth of Synthetic Biology(SynBio) led to signicant breakthroughs. CRISPR/CAS and the synthesis of synthetic chromosomes are just examples of what researchers achieved in the field in the last decade. However, the synthesis of synthetic constructs or organism is still a non-trivial and challenging task. To assure reproducibility and simplify the process, researchers rely on the Design-Build-Test and Learn cycle (DBTL), a work flow deriving from engineering disciplines. Therefore, engineers widely employ DBTL in Electronic Design Automation (EDA) and Computer-Aided Design (CAD), particularly for the design of circuits. As in CAD, researchers can select the components for a specific design and predict the outcome. Nevertheless, the biological nature of DNA components complicates the DBTL. DBTL consists of fours different steps: Design, Build, Test and Learn. The Design step consists of selecting the components of a synthetic construct (e.g. Promoters, Coding Sequences and terminators) and design the theoretical circuit. In the Build step, DNA synthesis and molecular assemblies allow to synthesise the components and assemble constructs. Then, the Verification step verifies that the sequence and the expression of the assembled constructs match with what specified in the Design step. The last step, Learn, evaluates the performances of the constructs and the analysed data informs the next Design step. The purpose of this work is to explore the use of the DBTL cycle in SynBio and introduce two methods focusing on different steps of the cycle: Multi-Objective Optimisation algorithm for the Design and Assembly of DNA constructs (MOODA) and Nanogate. MOODA is the rst genetic algorithm implemented for the Design and Assembly of synthetic constructs. One of the main limitations of DBTL is the dependency between steps, particularly between the Design and Build steps. The phosphoramidite synthesis imposes limits on the content and size of the molecule. As a consequence, researchers progressively edit the design to meet DNA synthesis constraints. As a result, the difference between the sequence of assembled and designed construct become significant. MOODA returns the list of optimised trade-offs between the Design and Build constraints. Nanogate instead derives from Nanopore and Golden Gate Assembly and focus on the Test step of the cycle since its purpose is to verify DNA sequences after assembly on an industrial scale. Nanogate takes advantage of Nanopore long reads since the length of a synthetic construct can usually be contained in a single Nanopore read. Nanogate, unlike standard alignment algorithms, employs hash codes to verify the similarity between the reads and the library of parts. The hash codes reduce the complexity of the algorithm and increase the True Positive rate.