Algorithms to Speed Up Drug Development
Developing algorithms to speed up a crucial aspect of biomedical research potentially shortening the time it takes for a drug to come to market.
Booz Allen/Kaggle, United States
Big data is not just a powerful and transformative tool for businesses. Information resulting from the analysis of zeroes and ones can deliver new solutions or provide new ways of understanding a vast range of problems from health issues to natural disasters.
Recognizing this, Kaggle, an online community for data scientists and management and IT consulting firm Booz Allen set up the Data Science Bowl in 2014. This annual open innovation competition brings together data scientists, technologists and experts from multiple industries to tackle pressing global challenges with data and technology.
Tackling Pressing Global Issues
“We created the Data Science Bowl in 2014 in order to convene communities to take on huge challenges, and give data scientists the opportunity to contribute to something bigger than themselves,” said Booz Allen Senior Vice President Josh Sullivan.
The fourth edition of the contest took place in 2018 and attracted nearly 18,000 global participants. They were tasked with automating the process of identifying cell nuclei in medical imagery to help speed up research for almost every disease. Each team had to create a computer model to identify a range of nuclei across varied conditions.
Speeding Up Medical Research
Identifying the nuclei of cells is important because it is the starting point for most analysis. Most of the cells in our bodies contain a nucleus, each one packed with DNA, the genetic code that programs each cell. Locating and recognizing the nuclei allows researchers to find each individual cell in a sample.
Measuring how cells react and respond to various treatments provides key insights into underlying biological processes. This knowledge is used as the basis for developing therapeutics. For example, a fault in a protein in a particular biochemical pathway that causes the symptoms of a disease can be corrected when the right drug targets specific protein receptors.
During the 90-day competition, more than 68,000 algorithms were submitted and evaluated. The overall winners were an international team from Russia and Germany. They had competed against each other in previous Data Science Bowls but decided to come together in 2018 with the expressed purpose of winning. For their efforts, they split cash and prizes worth $170,000.
“These solutions represent a paradigm shift in the way microscopy images are processed in biomedical research and will make research more accurate and efficient,” said Dr. Anne Carpenter, director of the Imaging Platform at Broad Institute of MIT and Harvard.
“The next step being investigated is to create a user-friendly and open-source software that biomedical researchers can begin using in their day-to-day work.”
Reducing Time-to-Market for New Drugs
Automating the process of nuclei detection will save researchers a lot of time, allowing them to focus on other aspects of their work. The hope is that in so doing the time taken for a drug to get to market will be considerably shortened. It currently takes about ten years before each new drug becomes available.
Next Story »