Deep Learning is a set of technologies which use huge amount of data to predict behaviors thanks to hidden (or not) signals inside. In this article for example, we see how a collaboration between Google scientists and Steve Finkbeiner’s use brain cells pictures to automatize “live / dead” classification, and many others ways to use it ! But, deep learning is not the “Holy Grail” of pictures interpretation, because algorithms has to be trained with many heterogeneous data, and it use also different methods. The challenge for biologists is to understand results and to link it with biology !

In HCS Pharma we currently produce huge amount of pictures and data and we are working on machine learning and deep learning technologies to help interpretation. First results are encouraging but we need time to propose a robust methodology to our customers. Stay connected !

More than 440 articles on the bioRxiv preprint server discuss deep learning; PubMed lists more than 700 references in 2017. And the tools are on the cusp of becoming widely available to biologists and clinical researchers. But researchers face challenges in understanding just what these algorithms are doing, and ensuring that they don’t lead users astray.

Source: Deep learning for biology

Deep Learning seems to be the magical word today in biology and medicine. This technology is used for DNA exploration, pictures interpretation or even to predict future diseases. Deep Genomics is going further and annonces to use of Deep Learning for early-stage development of drugs for Mendelian disorders. As the founder, Brendan Frey, says, it could be a new step in drugs developement, and maybe a “really massive shake-up of pharmaceuticals” !

Deep Genomics, a Canadian company that uses machine learning to trace potential genetic causes for disease, announced Tuesday that it’s getting into drug development. It joins a growing list of AI companies betting that their techniques can help produce powerful new drugs by finding subtle signals in huge quantities of genomic data.

Source: An AI-Driven Genomics Company Is Turning to Drugs – MIT Technology Review

Le diagnostique du cancer du poumon à l’aide d’images 3D à haute résolution est particulièrement efficace pour détecter les phases précoces de la maladie mais engendre encore trop de faux positifs.  Le Deep Learning peut aider à limiter cela en analysant un grand nombres d’images et en alimentant les algorithmes de détections avec des analyses validées. Pour améliorer ces algorithmes, les entreprises Kaggle et Booz Allen Hamilton ont fait de ce sujet l’objectif du Data Science Bowl de cette année.

Les résultats de cette compétition ont été très encourageants (voir l’article) et nous montrent encore une fois que l’utilisation d’une assistance numérique à la catégorisation et à l’interprétation des images en biologie n’est pas une option, mais une obligation dans les années qui viennent.

Depuis janvier, près de 10 000 data scientists du monde entier ont participé à l’Annual Data Science Bowl. Pour cette troisième édition, le challenge des analystes consistait à développer les meilleurs algorithmes d’apprentissage machine permettant d’aider les professionnels de santé à dépister plus tôt et plus précisément les cancers du poumon lors des examens tomodensitométriques.

Source: LeMondeInformatique

It’s very interesting to show that, at HCS Pharma, we work on many of the most exciting medical technologies identified by Dr. Bertalan Mesko on his blog, Of course, precision medicine is the most important for us, because we see here extraordinary tools to treat cancer and neurodegenerative diseases (“instead of canons, we start using sniper rifles.”)

But two other fields seem really promising. The 3D bioprinting for exemple will give us the possibility to create relevant in vitro models of living tissues. We also work to integrate IA deep learning in our HCS processes. It will be a major method to find phenotypic actions of futur drugs.

2016 was a rich year for medical technology. Virtual Reality. Augmented Reality. Smart algorithms analysing wearable data. Amazing technologies arrived in our lives and on the market almost every day. And it will not stop in the coming year.

Source: The Most Exciting Medical Technologies of 2017 – The Medical Futurist