The National Institutes of Health (NIH) has awarded FIU researchers a $1 million grant to design and develop machine-learning algorithms that allow biologists to make sense of proteomics, or the large-scale study of proteins.
The study of proteins is critical for understanding and treating diseases. In the past, biologists have spent long periods of time – sometimes their entire lives – studying as few as one single protein. In recent years, mass spectrometers, machines that use electricity and magnetism to analyze atoms, have allowed biologists to generate massive amounts of data from a few samples. The challenge is the data is difficult to understand with the naked eye.
With this three-year grant, Fahad Saeed, principal investigator and associate professor in the School of Computing and Information Sciences (SCIS) within the College of Engineering & Computing, along with other FIU researchers plan to develop machine-learning algorithms that work with supercomputers to analyze and understand the data. Supercomputers are large computers used for complex mathematical calculations, with speeds that greatly exceed that of standard computers.
COVID-19 is an example of an area where Saeed’s machine-learning algorithms can be used.
“The reason we don’t have a vaccine yet is that we don’t know enough about the virus,” Saeed said. “With the ability to study more data, quicker and more efficiently, biologists can learn more about these types of viruses and work towards a vaccine.”
Saeed and researchers propose novel artificial intelligence (AI) models that will allow biologists to make sense of the mass spectrometry data sets much more accurately compared to traditional numerical algorithms that have been used for the past 20 years. The idea is to study millions of proteins at once.
One thing Saeed and his team are keeping in mind is cost. They want these computational solutions to be available to small labs and developing countries that may not have accessible supercomputers.
Saeed is actively working toward finding FIU graduate and postdoctoral students. Part of the funding is to help fund students involved in this research.
The code Saeed is developing will be available on his lab’s website, SaeedLab - High-Performance Data Computing for Advancing Science & Society. The researchers are looking to create a service to help people download the code. The models themselves run on supercomputers.
The co-principal investigators of the grant are Shu-Ching Chen, associate director and professor in the SCIS; Jason Liu, professor and graduate program director in SCIS; Francisco Alberto Fernandez-Lima, associate professor in the College of Arts, Sciences & Education (CASE); and SS Iyengar, distinguished university professor and director of SCIS.
Additional reporting by Diana Hernandez-Alende