A The Kids Research Institute Australia researcher whose work as a young PhD student more than 20 years ago has helped to crack one of biology’s biggest mysteries has received an international peer award for making his work available to the open-source community.
Scientists have long sought to understand how to predict the structure of proteins made by all living organisms – critical knowledge that could help to supercharge research into a host of diseases.
Proteins and the chains of amino acids they are made from are the essential building blocks of living organisms, existing in every cell in the human body. A protein’s shape helps determine its function in the human body but until recently, scientists have only been able to measure the shape of a tiny fraction of the estimated 20,000 proteins expressed by humans.
In 2021, with a little help from an algorithm developed and made freely accessible to scientists everywhere by computational biologist Dr Timo Lassmann more than two decades ago, DeepMind largely solved this problem.
Now part of Google’s AI arm, DeepMind used the Kalign program created by Dr Lassmann to help build a breakthrough artificial intelligence method known as AlphaFold.
AlphaFold was subsequently used to accurately predict the shape of thousands of proteins, achieving within days what might otherwise have taken years to establish through traditional laboratory methods.
The work has produced what DeepMind co-founder and chief executive Dr Demis Hassabis has described as ‘the most complete and accurate picture of the human proteome to date’. The breakthrough is expected to accelerate the discovery of new drugs to treat multiple diseases.
Dr Lassmann, now an internationally respected leader in the field of computational biology, said he had been pleasantly surprised to discover that the DeepMind team had incorporated Kalign – albeit in a supporting role – into their ground-breaking AlphaFold method.
“This is particularly gratifying, as I have been an ardent admirer of the DeepMind team since their pioneering AI work on Q-learning and AlphaGo,” Dr Lassmann said.
Dr Lassmann has now received Google’s Open Source Peer Award for his work on Kalign – a project that began 23 years ago when, as a first-year PhD student with the Karolinksa Institute in Sweden, he was seeking to understand the evolutionary history and function of protein and DNA sequences.
Confronted with the daunting task of manually analysing thousands of sequences, he opted to create an algorithm to automate the process.
“Initially, I underestimated the complexity of this task, but it forced me to deeply study the analysis of biological sequences and master advanced programming techniques, leading to a significant breakthrough,” Dr Lassmann said.
The work culminated in six first-author PhD publications, with Dr. Lassmann releasing all his computer codes as open-source software – making them freely accessible to the global academic community.
“Fortuitously, the utility of Kalign grew due to its ability to scale to extremely large projects powered by the development of genome sequencing technologies,” Dr Lassmann said.
Kalign is now being used across a wide range of applications, from constructing the human genome reference annotation to studying the mutational processes in SARS-Cov-2 and antimicrobial resistance genes, and virulence factors in bacteria.
Dr Lassmann said the achievement demonstrated the far-reaching influence of basic discovery science.
“Discovery science is not merely a precursor to translational research – it is the principal driver of scientific innovation,” he said.