Indian-origin US Entrepreneur’s AI Algorithm Can Spot Healthiest Sperm

To develop the sperm health test algorithm, the team from California-based Oma Fertility Clinic scanned sperm from the ejaculates of more than 1,000 men.

New Delhi: An Indian-origin US entrepreneur’s company is working on a novel Artificial Intelligence (AI)-based algorithm that can identify the healthiest sperm, an advance that can reduce the time and costs of in-vitro fertilisation (IVF) pregnancies.

To develop the sperm health test algorithm, the team from California-based Oma Fertility Clinic scanned sperm from the ejaculates of more than 1,000 men, Daily Mail reported.

The AI algorithm scans sperms for their shape and how rapidly they move to pick out the best ones for fertilising an egg — a process the company believes will boost success rates with IVF and lower costs, as couples will need fewer rounds of the treatment.

“As it stands right now, couples have to go through three cycles of IVF on average,” Dr Kiran Joshi, co-founder at Oma Robotics, was quoted as saying to DailyMail.

“We want to reduce that number as that will reduce the pain and reduce the financial burden and lead to more success,” he said.

According to studies, about 40 per cent of infertility cases are because of the male partner having low sperm counts — majorly on the rise due to increase in unhealthy diets and sedentary lifestyles.

The sperm samples for the algorithm were mostly from men in their 30s and 40s, from across the US and other countries. Some also ranged up to the age of 75 years.

Dr Joshi’s team extracted and analysed a subset of sperm from each sample — which contained up to 20,000 sperms, based on their shape and how well they swim — an exercise which Dr Joshi called as “like searching for a needle in a haystack”, the report said.

A healthy sperm has a smooth and oval-shaped head, and will also be able to swim rapidly and in a straight line.

The sperm were graded based on these two aspects using standardised measures and an about 10 second video of each was then recorded, which was then fed into the AI to give it the ability to accurately estimate how healthy any one particular sperm is.

The AI algorithm delivers results within seconds. It draws a box around each sperm and then quickly allocates a colour based on the perceived health of the sperm. The red boxes indicate unhealthy sperm, while green indicate healthy, the report said.

The findings, which show a ‘significant’ difference between the two groups, will soon be published in an academic journal, the team said.