Image Recognition in the Hunt for New Natural Products

We are in desperate need of new antibiotics, but finding new natural products with pharmaceutical potential by screening isolated bacteria is by definition a problem of finding the needle in the haystack. By using image recognition and machine learning, we want to develop an incubation chamber with a build-in software capable of monitoring agar plates of growing bacteria and pinpoint the colonies that are most likely to produce novel natural products. This could have an enormous impact on the natural product field by providing a powerful tool for prioritizing drug-discovery efforts. Moreover, it could allow screening in remote and underdeveloped countries that have rich biodiversity, but little access to high-end analytical equipment otherwise used in dereplication.