New Machine Learning Model Simplifies Complexity of Gene Regulation
AI and machine-learning algorithms are increasingly allowing biologists to get better insights into human diseases. Unarguably, molecular pathways underlying gene regulations by cells have been treasure-trove of information in therapeutics. However, growing complexity of this mechanistic knowledge has been a key cause for concern for biologists. With time, these AI algorithms have become difficult to understand, and harder to apply. Further, they are not how biologists would think.
Making Artificial Neural Network Easier to Understand for Biologists
Meanwhile, two quantitative biologists sought to make advanced machine learning algorithms easier to understand. They developed artificial neural network (ANN) to represent mathematical thermodynamic model for gene regulation. Further, ANN is particularly useful in investigating DNA through a process called ‘massively parallel reporter assay.
The uniqueness of the machine learning model, the duo contend, is that it is fashioned in the way biologists would ask questions. In addition, this would facilitate their efforts in exploiting the assay data for disease modelling. This would be a game-changing application of AI, advancing the research pertaining to molecular therapies against diseases.
Model Facilitate Molecular Therapies for Human Disease Modelling
As the name suggests, scientists have made ANN on the lines of how neurons connect and branch in the human brain. Interestingly, the new model can be key computational tool for cutting-edge machine learning platforms used in life sciences industry. Particularly, their application will help biologists develop molecular therapies faster and with more precision.
The researchers presented their findings at the 1st Conference on Machine Learning in Computational Biology recently. The machine learning model can become a promising candidate for studying key gene circuits.