Microservices for collaborative research

Cells are being genetically modified in labs all over the world for different purposes: from curing diseases to making ecological biofuels. Scientists are predicting cells behaviour using all kinds of computer models. There is a lot of accumulated knowledge in the field, but lack of compatibility and reusability of software is a huge obstacle on the way to greater science. Microservices concept fits naturally to the task of connecting and scaling software built by research teams with different backgrounds spread across different countries. We will talk about, what kind of software modern biology has and needs, how microservices architecture can boost up scientific collaboration and what are the challenges on the way.

Transcript

1. Microservices for collaborative research
2. The Novo Nordisk Foundation Center for Biosustainability cell factories design, construction and deployment
3. What is a “cell factory”?
4. What is a “cell factory”? • Cells are designed by evolution to produce certain chemicals • It is possible to change their fermentation behaviour by changing the genome • There are computer and mathematical models for doing it smarter
5. DD-DeCaF (Data-Driven Design of Cell Factories and Communities)
6. Use case: reactions knockouts
7. The microservices approach
8. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
9. Microservices for academia • Independent and parallel development • Small, easy to understand code base • Different technology stacks • Easy come, easy go • Bounded contexts
10. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
11. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
12. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
13. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
14. Microservices for academia • Independent and parallel development • Separate, easy to understand code base • Different technology stacks • Easy come, easy go • Reusability • Immutable infrastructure
15. • Easy switch between server implementations using the same message objects frontend another service HTTP RPC https://github.com/biosustain/venom https://github.com/lyschoening
16. • Currently available: ◦ Unary requests & responses with ▪ aiohttp ▪ grpcio • Coming soon: ◦ Streaming requests & responses with ZMQ ◦ Unary HTTP/1 implementation with Flask ◦ WebSocket implementation with aiohttp ◦ gRPC (HTTP/2) implementation with e.g. hyper-h2 ◦ MsgPack serialization support ◦ Protocol Buffer support ◦ JSON Hyper-Schema service ◦ OpenAPI (Swagger) schema service ◦ Client generation from schema https://github.com/biosustain/venom https://github.com/lyschoening
17. • The research institute official website ◦ http://www.biosustain.dtu.dk/ • The project ◦ http://dd-decaf.eu/ ◦ @DDDeCaF • Github ◦ https://github.com/biosustain ◦ https://github.com/dd-decaf Links and acknowledgements
18. We’re hiring! • 15 new computational people to be hired at Technical University of Denmark - half software/infrastructure, half scientific • Expanded activities include: ◦ Maintenance and curation of genome-scale models ◦ Structural bioinformatics ◦ Big data analysis ◦ Laboratory information management system and design software development http://www.biosustain.dtu.dk/ svegal@biosustain.dtu.dk