Scientists have created a free software and web database aimed at overcoming the challenges of identifying critical protein-protein interactions that can be targeted for medication. This computational tool is named PIONEER, which stands for Protein-protein InteractiOn iNtErfacE pRediction. Researchers showcased how useful PIONEER is by pinpointing potential drug targets for various cancers and other complicated diseases.
Scientists from Cleveland Clinic and Cornell University have developed a free software and web database to help identify important protein-protein interactions for medication treatment.
The computational tool, called PIONEER (Protein-protein InteractiOn iNtErfacE pRediction), has been demonstrated by researchers to identify prospective drug targets for numerous forms of cancer and other complex illnesses, as detailed in a recent Nature Biotechnology publication.
According to Feixiong Cheng, PhD, the co-lead author of the study and director at Cleveland Clinic’s Genome Center, genomic research is crucial for drug discovery, but it’s not always sufficient by itself. It typically takes around 10-15 years to move from discovering a gene responsible for a disease to launching clinical trials for a drug based on that genomic data.
“In principle, creating new medications using genetic data seems simple: mutated genes produce mutated proteins,” Dr. Cheng explains. “We aim to develop molecules that prevent these mutated proteins from interfering with essential biological processes by stopping them from interacting with healthy proteins, but in practice, achieving this is far more complicated.”
Each protein in our body can engage with numerous other proteins in various ways. These proteins can further interact with hundreds more, creating a complex network of interactions known as the interactome, according to Dr. Cheng. The situation becomes even more intricate when you include disease-causing DNA mutations. Some genes can be altered in numerous ways, leading to the same disease, meaning a single condition can link to multiple interactomes arising from just one differently mutated protein.
This leaves drug developers with countless potential disease-causing interactions to consider. This is only after they have compiled a list based on the physical structures of the affected proteins.
Dr. Cheng aimed to develop an artificial intelligence (AI) tool to aid genetic/genomic researchers and drug developers in easily identifying the most promising protein-protein interactions. He collaborated with Haiyuan Yu, PhD, who leads the Cornell University Center for Innovative Proteomics. Together, they combined extensive data from various sources, including:
- Genomic sequences from nearly 100,000 individuals who either had disease-causing mutations from birth or developed them later in life, usually due to cancer.
- Three-dimensional structures of over 16,000 human proteins, along with data detailing how DNA mutations affect these structures.
- Documented interactions among nearly 300,000 different protein-protein pairs.
The resulting database enables researchers to explore the interactome for more than 10,500 diseases, ranging from alopecia to von Willebrand Disease.
Researchers who find a mutation linked to a disease can input it into PIONEER to receive a ranked list of protein-protein interactions contributing to the disease that could potentially be targeted with a drug. Scientists can also look up a disease by its name to obtain a list of likely disease-causing protein interactions for further investigation. PIONEER is crafted to support biomedical researchers across various diseases, including autoimmune disorders, cancer, cardiovascular conditions, metabolic issues, neurological diseases, and pulmonary ailments.
The research team confirmed their database’s predictions in laboratory tests, where they created nearly 3,000 mutations across over 1,000 proteins and examined their effects on almost 7,000 protein-protein interaction pairs. Initial studies based on these results are already in progress to develop and test treatments for lung and endometrial cancers. Furthermore, they demonstrated that the model’s predicted protein-protein interaction mutations could forecast:
- Survival rates and prognoses for various types of cancer, including sarcoma, a rare but potentially lethal cancer.
- Responses to anti-cancer drugs as observed in large pharmacogenomics databases.
Additionally, the researchers confirmed that mutations affecting protein-protein interactions between NRF2 and KEAP1 can indicate tumor progression in lung cancer, presenting a new target for developing specific cancer treatments.
“The resources required for interactome studies present a significant barrier for many genetic researchers,” Dr. Cheng said. “We hope PIONEER can help alleviate these challenges through computational means, empowering more scientists to advance new therapeutic options.”
