[Studentship] AI-driven design of enzyme replacement therapies
Ente: Other NPIF
Scadenza: 2023-08-30
Paese: GB
Descrizione
Enzymes are molecules catalysing all the reaction in a cell and, when defective, they can cause life threatening diseases, such as Gaucher's disease. Patients with enzymatic deficiencies can be treated by injecting a recombinant version of the defective enzyme; however, injection of wild-type human enzymes is usually ineffective, as they are less active and often cause immune response. Thus, human enzymes must be engineered to optimise their therapeutic properties.
Designing enzymes is challenging as standard biochemical or biophysical methods do not work well on large proteins as they are not sufficiently accurate to identify new functional enzymes in large protein sequence space. Generative Machine Learning (ML), instead, represents an attractive approach to learn design principles to build new enzymes directly from the sequences of molecules found in nature. Recently, conditional recurrent neural networks (cRNN), a particular type of neural networks (NN) suitable for working with sequences, were successfully used to generate novel antimicrobial peptides and protein structures, but scaling these methods to design complex molecules, including enzymes, remains a largely unexplored field.
We will develop deep generative models to learn the functional design space of human enzymes and implement optimisation methods to find the most likely amino acid sequence encoding a specific catalytic function. To do that, we will integrate proteomic, evolutionary and structural data publicly available across biological databases and develop variational methods to fit enzyme sequence models. The availability of these methods will enable design of new designer enzymes.
Settori: Sch of Informatics
Vai al bando originale
Registrati gratis su Bandolo per trovare bandi compatibili con la tua azienda.