A global team of researchers led by Nigerian scientist Dr Elijah Kolawole Oladipo has used artificial intelligence to find interesting compounds from animals that could lead to better focused breast cancer treatments with fewer side effects. Africans & Diaspora
Published in the Future Journal of Pharmaceutical Sciences, the research combined AI-enabled drug discovery and sophisticated 3D protein modelling to identify therapeutic peptides that, in computer simulations, might target proteins driving breast cancer more effectively than an existing FDA-approved anti-cancer peptide.
According to Saturday Guardian, breast cancer is still one of the deadliest diseases among women globally, representing over 36 per cent of cancer diagnoses in women and causing millions of new cases each year.
Conventional treatments such as chemotherapy, surgery and radiation sometimes harm good cells as well as diseased ones, and many cancers eventually become resistant to existing medications.
To tackle the challenge, Oladipo of Adeleke University and the Helix Biogen Institute headed teams of researchers from Nigeria, Ethiopia, the United Kingdom and the United States in designing potential cancer-fighting molecules using an AI-based approach rather than depending on years of laboratory testing.
The international team included Omolara Omoboye Adegboye, Stephen Feranmi Adeyemo and Modinat Wuraola Akinboade of Helix Biogen Institute; Prof. Bamidele Abiodun Iwalokun, of the Nigerian Institute of Medical Research; Dr. Olumide Faith Ajani, Africa Centers for Disease Control and Prevention, Ethiopia; Dr. Olumuyiwa Elijah Ariyo, Afe Babalola University and Federal Teaching Hospital Ado-Ekiti; Prof. Helen Onyeaka, University of Birmingham; and academic collaborators at Stony Brook University, New York. Africans & Diaspora
Speaking on the importance of the research, Oladipo said the project showed how AI may drastically accelerate the search for better cancer medicines.
“We turned to artificial intelligence to design precision therapeutics inspired by nature to create therapies that can attack cancer cells while sparing healthy tissue,” he stated.
The researchers used machine-learning algorithms and Colab AlphaFold2, a cutting-edge protein-structure prediction technology, to scan a global database of over 1,500 natural peptides rather than testing thousands of chemicals in the lab.
The AI algorithm screened the chemicals for structural stability, toxicity and allergenic potential, before reducing the search to three top possibilities from animal sources.
The nominated compounds were then virtually simulated against two proteins essential for breast cancer progression – Matrix Metalloproteinase 1 (MMP1) which allows cancer to spread and Epidermal Growth Factor Receptor (EGFR) which promotes tumour growth.
One AI-optimised peptide, Metalnikowin IIA, based on the green shield bug, demonstrated improved binding affinity, increased accuracy and enhanced structural stability compared to Buserelin, an FDA-approved anticancer peptide that was utilised as a benchmark during the simulations.
The researchers did caution however that the findings are still at the computational stage and have not been verified in the lab or clinical tests yet.
“The findings are at the AI-predicted stage at this point,” the researchers said, noting that there’s a need for further testing in live cells and animal models before the chemicals can be considered for treatment of humans.
The team is seeking international research grants, pharmaceutical partnerships, and global health collaborations to fund lab validation and take the research beyond computer simulations.
This funding would provide the critical link between digital innovation and real-world clinical cancer care, which could open a new chapter in the advancement of more accurate and less dangerous breast cancer treatments, the researchers added.
About the Author
