University of Glasgow Researchers Propose A Smartphone-Based DNA Diagnostics For Malaria Detection Using deep Learning

Malaria is one of the world’s leading reasons for illness and death. Moreover, malaria can spread asymptomatically, making widespread field testing important to contain outbreaks and save lives.

A significant challenge to make field testing largely available is that the most accurate and standard method for the malaria blood test is based on the polymerase chain reaction (PCR) process. PCR tests usually require trained staff to draw blood and laboratory conditions to test the samples. In remote areas such as sub-Saharan Africa, malaria infections usually break out miles away from trained staff and lab conditions, thus making effective infection control quite tricky.

Over the past few years, a team of biomedical engineers from the University of Glasgow, in collaboration with the Ministry of Health in Uganda, has developed a more low-cost, reliable, ‘origami’ alternative to PCR and lateral flow tests.

The method uses sheets of folded wax paper to prepare patient samples for a different detection process known as loop-mediated isothermal amplification (LAMP), which can be delivered in the field. Field tests in Uganda have demonstrated that the origami test technique is 98% accurate.

The blood sample taken from a patient via fingerprick is placed in a wax channel on the folded paper. Next, the paper is folded, directing the sample into a narrow channel and then to three small chambers, which the LAMP machine uses to test the sample’s DNA for evidence of Plasmodium falciparum, the mosquito-borne parasitic species responsible for causing malaria.

The researchers also describe how they developed a secure smartphone app to pair with their origami tests, which uses deep learning to facilitate more accurate diagnosis and allow better community transmission surveillance.

The LAMP results are analyzed utilizing a cloud-based machine learning process to ensure that they are correctly administered, allowing users of varying skill levels to conduct the test correctly. A negative or positive diagnosis of the patient’s malaria infection is provided using lines on a lateral flow strip synonymous with those used for home COVID-19 testing. The patient’s results are stored on a blockchain-based ledger to ensure privacy and shared with the local authorities to allow anonymized monitoring.

Professor Jon Cooper, who led the development of the diagnostic system, believes that the system they have developed could help deliver on a few urgent requirements. It allows non-experts to administer blood tests anywhere and then securely share those results with local and regional authorities.

The researchers validated their method using the help of some initial field tests in Uganda’s rural Eastern Tororo District. They utilized their system to test blood samples collected from 40 school children from a local primary school-aged. All samples were then re-tested in the UK using a standard PCR test for malaria. The tests were found to be 98% accurate.

According to Dr Julien Reboud, another co-author of the paper, smartphones are widely used in Africa, thus making them invaluable in enabling widespread testing and effective surveillance of infectious diseases like malaria.

This new digital health technology uses artificial intelligence and blockchain on a mobile phone to deliver field-based DNA testing along with an expert decision diagnostic support tool.

The testing in Eastern Uganda highlighted the high local burden of disease; it allowed the researchers to treat affected children timely and precisely and enabled them to provide data to regional health and education authorities to adjust broader disease management strategies locally. The team’s paper is published in Nature Electronics.



🐝 Join the Fastest Growing AI Research Newsletter Read by Researchers from Google + NVIDIA + Meta + Stanford + MIT + Microsoft and many others...