A short-cut to discovering new medicine
To produce first-in-class medications, drug discovery and development experts have used a standard process for decades. Understanding pathogenesis, selecting targets, developing bioassays and subsequent high throughput screens are the typical first four steps of a first-in-class discovery effort. These stages can take over three years to complete. Today, companies are emerging that replace these four processes which run completely in silico and produce the same output (a pre-clinical ready molecule) in less than a month.
Using AI algorithms to detect breast cancer
Last month, a team of researchers announced that they had created an AI algorithm which outperformed radiologists in detecting breast cancer in mammograms. Trained on tens of thousands of mammogram images, the algorithm measurably reduced both false positives and false negatives compared to human doctors. Currently, the consensus is that AI will not replace radiologists, but radiologists using AI will replace those who do not.
Embrace the data
In recent years we have witnessed large tech companies disrupt industries, one after the other, all by using data. Take Uber for example - they’re the world’s largest car hailing company and they do not even own a single car! The same disruption is coming to the life science field. We will undoubtedly see a world where data driven science, the use of AI, and deep learning will play a major role. It will be a world where development of pharmaceuticals happens in a lab full of super computers and data scientists - not test tubes and petri dishes.