This PhD student has built a chip to detect diseases earlier through AI
The first time Newcastle University PhD student Daniel Todd set out to build a chip from scratch that could help scientists detect diseases early, he used a couple of Tupperware containers and small electric motors typically used by hobbyists for flying model planes to create his first mould.
In chip manufacturing, moulds are created through photolithography, a complex and traditionally expensive process that involves applying a thin layer of light-sensitive liquid over a silicon wafer evenly to transfer tiny patterns through a mask with UV light — something akin to developing film in photography.
It took Todd 17 failed attempts before getting it right every time with homemade tools, replicating a process that typically requires billions of dollars of investment for less than £200.
“I’ll often start from a position of childlike imagination,” Todd says. “I go as fundamentally first-principles as I can. In this case, I needed a way of capturing small amounts of very small molecules simply and quickly, so I needed loads of holes to stick these in, and I started thinking about it in terms of billiards or snooker.”
Eventually, he built a series of functioning microfluidic chips — interfaces that combine biology and hardware with important applications across the future of computing — patterned with up to 10 million of tiny holes that are invisible to the naked eye, each the size of a red blood cell, which could be used to test samples for the presence of foreign DNA, detecting viruses, pathogens, tumours and other diseases early.
The identification of target materials in samples is then automated through an AI-powered scanner capable of detecting them in real time, including their nature and concentration, and returning results within seconds to any device.
Todd used his first chips to isolate and detect salmonella and gonorrhoea, but he says the ability to scale up across different industries is enormous — “from preventing infections from ruining crops to helping scientists cure cancer” — and he has recently launched a startup, InvenireX, to do just that.
The goal is to initially provide researchers with tools that can speed up discoveries effectively before expanding to other areas of diagnostics. The startup has also secured a collaboration with the Food and Environment Research Agency to start testing whether its technology can detect blight in leaves early enough to save the infected plant.
InvenireX is one of the teams on the 2023 Conception X Northern Deeptech Hub cohort. Before joining the deeptech venture programme for PhD students, Todd had always known he wanted to invent things and experiment with science beyond academia.
“I was always aware that in order to create change, I would have to commercialise my research,” he says. “Having the necessary technological proficiency is one thing, but moving from researcher to entrepreneur is something that would be extraordinarily difficult to do without support.”
“Conception X has provided a map that I can follow, introduced me to VCs and other founders, and helped me find my footing into this new world through coaching.”
Five months into the programme, InvenireX has grown into a team of four, built and refined its prototype, moved into its first-ever office, and conducted dozens of conversations with potential customers. It has also received £100k in funding through Conception X’s partnership with XTX Ventures, and plans to close a £700k round by the end of the year.