Advanced robotic micro-manipulation of gametes

Infertility is a global concern, affecting approximatively 1 in 6 couples who are trying to conceive. In vitro fertilisation (IVF) is an effective fertility treatment; however, worldwide only a privileged 2% of the clinically infertile can access IVF. In the UK, approximatively 60% of the couple who could benefit from IVF still do not have access to treatment and most UK patients end up paying out of pocket due to limited public funds. The high cost of treatment, together with the low success rate of around 33%, prohibit the widespread use of IVF.

The proposed project is part of a collaboration between King’s College London and Conceivable Life Sciences, partners who co-fund the research. Conceivable aims to revolutionise IVF by providing end-to-end automation of the IVF lab. A radical approach wherein the full IVF pipeline is being reconsidered and modernised with the help of robotics and artificial intelligence is undertaken.

The proposed PhD project relates to the advanced micro-manipulation of gametes, i.e. sperm and oocyte, towards the goal of robotic intracytoplasmatic sperm injection (ICSI). Specifically, research will focus on the following key areas:

Sperm identification, tracking, and loading

The student will start by researching computer vision methods to detect and track motile sperms under optical microscopy. Artificial intelligence methods that enable microstructure tracking in real-time will be considered, with emphasis on tracking the tail motion characteristics of sperm to select the most preferred for insemination. The project will then use a visual-servo control approach of a laser head to immobilise sperm, prior to it being aspired in a robotically actuated micropipette. The dynamics of sperm loading based on fluid viscosity, micropipette characteristics, orientation, etc will be studied, and a reinforcement learning-based controller will be derived.

Oocyte identification, tracking, and manipulation

To increase fertilisation success rates, oocytes need to be preferentially aligned prior to insemination. This implies understanding of oocyte morphological characteristics under optical microscopy, e.g. the polar body which needs to be spared during insemination. Deep learning approaches will semantically interpret oocyte images. Alignment and re-orientation of the oocyte will happen through micro-suction cups created through the project, or through the generation of microfluidic vortices that manipulate the egg in a non-contact fashion.

Insemination

Gripe-needle technology will be miniaturised to allow both for oocyte manipulation, and sperm injection exploiting suction as a constraining factor. Micro/nano force sensors and contour analysis will provide data that guide the safe and effective insemination, gathering downstream evidence of success rates. Innovation in insemination approaches will be key to the success of the project, needing to explore 3D OCT imaging, horizontal vs vertical injection, variety of speeds, and whether model-predictive control or reinforcement learning can be applied.

Informal email enquiries from interested students to the supervisor are encouraged (contact details below):

Prof. Christos Bergeles, [Email Address Removed]

Further information about Conceivable can be found here: https://www.conceivable.life

For further details on eligibility and the application process please visit the studentship webpage.