plantCODE:

Recoding plants


Humans have been manipulating plant form by selection and breeding for over 12,000 years, to generate crop varieties with stable, engineered new morphologies. These domesticated plants form the bedrock of modern agriculture.

Many of the genetic changes associated with crop domestication have been mapped precisely. The domestication of crop plants like corn or tomato is associated with genetic changes of limited number and type - and are generally changes that trigger reconfiguration of regulatory networks.

New gene editing techniques have allowed scientists to recapitulate these traits in wild plants. This introduces the prospect of being able to systematically reprogram the growth and final form of any plant, and being able to harness the functional diversity of the ~400,000 plant species that have not been domesticated. For this, we need to understand the relationships between DNA code, cellular code and plant form.

 Tomato varieties. Image: Harry Klee, University of Florida

Tomato varieties. Image: Harry Klee, University of Florida

See:


Marchantia polymorpha is a lower plant system with (i) simple genome and architecture and (ii) ease of culture and genetic manipulation, and allows (iii) extraordinarily facile visualisation of genetic, cellular and physical dynamics using quantitative live microscopy.

Machine learning systems, inspired by biological neural networks, allow new approaches to analysis of complex systems. Cell-specific genomic datasets and high throughput microscopy surveys of cellular dynamics provide training sets for AI algorithms, in order to capture executable rules for organism-wide genetic and cellular interactions.

The phytoCODE: research unit brings together a group of experts in Marchantia biology, synthetic biology, gene editing, cell specification and genome-wide gene expression, quantitative microscopy, image processing, machine learning and model building for design and engineering. 

Neural network algorithms provide new and efficient methods for recognising and classifying cellular features, and decoding complex interactions between genetic elements, and between local cell populations.

 Visualising cell proliferation during Marchantia growth. Image: Nuri Purswani, Haseloff Lab

Visualising cell proliferation during Marchantia growth. Image: Nuri Purswani, Haseloff Lab

Grey0.png

Learn:


Marchantia tissues have an extreme propensity to regenerate after transformation. It also spontaneously produces clonal propagules. Engineered plant lines can be screened within a few weeks of an experiment. Patterns of gene regulation and cellular dynamics can be directly analysed in these propagules (right).

AI-based software models will be simple to explore and validate with this experimental system, where key parameters are directly visible. They will provide a test bed for generalised approaches to neo-domestication by gene-editing. Executable models have the potential to allow in silico evolution of new genetic blueprints, and design of prototype traits for future crop plants.

We hope that this work will contribute to the establishment of a centre for the application of AI models in whole-organism bioengineering, and support innovative approaches to biodesgn for agriculture and future sustainable industries.

 Live imaging of gene regulation in growing Marchantia. Image: Bernardo Pollak, Haseloff Lab

Live imaging of gene regulation in growing Marchantia. Image: Bernardo Pollak, Haseloff Lab

Build:


Machine learning techniques provide new and effective approaches to simplifying complex systems and the building of executable models. New classes of neural network algorithms allow the construction of models with exposed features that can correspond to known genetic parameters. Such software models for large-scale cellular growth will provide an instruction manual for rewiring genetic and cell-interactions where, for the first time, there is way of modelling the phenotypic consequences of engineered changes in the DNA.

Marchantia contains most of the biological machinery found in other terrestrial plants, but often in starkly reduced form. For example, many transcription factors, and important components of cell signalling are found as single gene copies. Marchantia provides an ideal testbed for building and validating new genetic modules for plant neomorphogenesis or physiological regulation. 

In Cambridge, we have built an extensive library of standardised DNA parts for Marchantia, have efficient automated systems for high throughput assembly of genetic circuits for plant transformation, and established methods for fast gene editing in the haploid organism, which can be used for testing models and engineering new plant species.

 Mining the Marchantia genome for standardised DNA parts. (http://marpodb.io)

Mining the Marchantia genome for standardised DNA parts. (http://marpodb.io)

 Rapid Loop assembly of DNA parts to build genetic circuits for reprogramming Marchantia. (Bernardo Pollak & Fernan Federici)

Rapid Loop assembly of DNA parts to build genetic circuits for reprogramming Marchantia.
(Bernardo Pollak & Fernan Federici)

Grey0.png

Share:


Current IP practices and restrictive licensing threaten to restrict innovation as the scale of DNA systems increases. We believe that the  field needs to explore new “two-tier” intellectual property models that will protect investment in applications, while promoting tool sharing, knowledge exchange and innovation in enterprises and developing bioeconomies, globally.

Cell-free systems provide a simple way to implement programmatic approaches to biology for both project-based teaching and field applications for diagnostics and environmental sensors. Cell-free extracts are GMO-free and can be distributed as stable freeze-dried powders for use where resources are limited.

We have sponsored a new legal framework for free exchange of DNA building blocks (OpenMTA) and promote development of project-based teaching and open curricula (Biomaker). We work with partners in Africa, Latin America and the Cambridge Global Challenges Initiative to promote exchange of knowledge and materials, internationally.

 http://www.openmta.org

http://www.openmta.org

 https://www.openplant.org

https://www.openplant.org

 https://www.biomaker.org

https://www.biomaker.org

Applications


Plants are platforms for bioproduction. Their modular and plastic body plans, capacity for photosynthesis, extensive secondary metabolism, and agronomic systems for gigatonne-scale production make them ideal targets for genetic reprogramming. However, efforts in this area have been constrained by slow growth, long lifecycles, the requirement for specialized facilities, a paucity of effcient tools for genetic manipulation and especially, the complexity of multicellularity. We have identified new experimental and theoretical frameworks for modelling the way plant genetic networks, cellular populations and tissue-wide physical processes interact at different scales.

We can construct DNA code and edit existing plant genomes.  AI-based multi-scale models for plant growth offer the prospect of rational design and programming of new plant traits, with potentially revolutionary consequences for agriculture and bioproduction. 

 Dividing and differentiating plant cells. (Fernan Federici, Haseloff Lab)

Dividing and differentiating plant cells. (Fernan Federici, Haseloff Lab)

Grey0.png

Future plans

Future_CODE_org_plan2018.png

CODE: Centre for Organismal Design and Engineering