Hey guys, I am thinking about a new project. I am bioinformatician in a small lab, where we have mostly WES data. I would like to learn machine learning programming, I am OK with writing scripts in python, make data analyses with pandas etc. A few months ago I made the Machine Learning course by google (https://developers.google.com/machine-learning/crash-course/ml-intro), now I have a plan for my project:

1) We have plenty of annotated WES variant with classification made (each variant has pathogenicity evaluated by annotations)

2) I will gather these annotated WES data and I will use a machine learning approach to train the model for evaluation of annotated WES data - it will help decide pathogenicity of the variant or at least helps to show which annotations are more significant.

I am planning to use TensorFlow and python.

Do you think, that this approach is good way to go, or I am completely "off"?