PhD's and Gordon Plotkin
I had a meeting this morning with Gordon Plotkin to discuss potential
PhD topics. It was brilliant. He gave me three options the best of
which I'm going to go on about!
A High-Level Language for Complex Signalling Pathways, applied to the Immune System
Biologists use many informal pictorial representations to express metabolic, signalling and regulatory pathways. However, these are not adequate for systematising current knowledge of very complex systems of pathways, e.g. the macrophage interferon signalling pathways in the immune system (What???? will need to learn about this!). Standardisation is therefore necessary. Starting from work of Kohn and Kitano, Ghazel et al have developed a standard set of graphical (pictorial) symbols, adequate to give a logical and contextual description of biological pathways. However, the resulting diagrams are very large and monolithic: there is no way to divide the system up into a collection of connected subsystems, or to link up the resulting static picture with a system dynamics where qualitative or quantative simulation studies can be carried out. It is also awkward to make versions of the model, or simplifications more amenable to analysis.
Computer science experience indicates the benefit of programming language approaches over pictorial ones. Programming languages can express modularity, allowing straightforward changes of system structure or components. Plotkin, building on recent computer science work by Regev, Priami and others, has developed a basic language to naturally express chemical reactions occuring in nested compartments; the language can also describe modified proteins, and their complexes. Programs can be translated to diagrams (Perti nets) easily comprehensible by biologists, which support both qualitative and quantitative dyamics.
The language is, however, too low level. The Ph.D student (make it Ianthe!!!) would build on the above work to design a high-level language whose primitives correspond to Ghazal's graphical ones. They would test the expressiveness of this language by writing down a modular description of the interferon pathway. They would write a translator between the programs and diagrams, on the one hand allowing pictorial representation of both modular structure and pathway details to be explored on screen, and, on the other hand, system components and their connections to be input on screen. This should be a particularly useful tool for biologists wishing to use the language.
I am so taken by the whole idea! It's a four year PhD as the first year will mostly be taken up learning the relevant Biology and Chemistry but I'm okay with that. It's so fascinating. Wow. I am so happy. He says that it is very competitive though and I would definately need to get a first in order to be accepted for it. Studying like a mad-ass about to commence.
A High-Level Language for Complex Signalling Pathways, applied to the Immune System
Biologists use many informal pictorial representations to express metabolic, signalling and regulatory pathways. However, these are not adequate for systematising current knowledge of very complex systems of pathways, e.g. the macrophage interferon signalling pathways in the immune system (What???? will need to learn about this!). Standardisation is therefore necessary. Starting from work of Kohn and Kitano, Ghazel et al have developed a standard set of graphical (pictorial) symbols, adequate to give a logical and contextual description of biological pathways. However, the resulting diagrams are very large and monolithic: there is no way to divide the system up into a collection of connected subsystems, or to link up the resulting static picture with a system dynamics where qualitative or quantative simulation studies can be carried out. It is also awkward to make versions of the model, or simplifications more amenable to analysis.
Computer science experience indicates the benefit of programming language approaches over pictorial ones. Programming languages can express modularity, allowing straightforward changes of system structure or components. Plotkin, building on recent computer science work by Regev, Priami and others, has developed a basic language to naturally express chemical reactions occuring in nested compartments; the language can also describe modified proteins, and their complexes. Programs can be translated to diagrams (Perti nets) easily comprehensible by biologists, which support both qualitative and quantitative dyamics.
The language is, however, too low level. The Ph.D student (make it Ianthe!!!) would build on the above work to design a high-level language whose primitives correspond to Ghazal's graphical ones. They would test the expressiveness of this language by writing down a modular description of the interferon pathway. They would write a translator between the programs and diagrams, on the one hand allowing pictorial representation of both modular structure and pathway details to be explored on screen, and, on the other hand, system components and their connections to be input on screen. This should be a particularly useful tool for biologists wishing to use the language.
I am so taken by the whole idea! It's a four year PhD as the first year will mostly be taken up learning the relevant Biology and Chemistry but I'm okay with that. It's so fascinating. Wow. I am so happy. He says that it is very competitive though and I would definately need to get a first in order to be accepted for it. Studying like a mad-ass about to commence.