Summary

A successful application of DL for a genomics problem heavily relies not only on developing an accurate model but also on how to make the model impactful. Model deployment is the process of transitioning a trained model built on notebooks into a production environment where it is used for prediction, classification, clustering, and other purposes. Unlike model training, deploying models requires different skills that are not traditionally taught to data scientists and other genomic scientists because these skills, such as web app development, cloud computing, and working with APIs, are more software development skills. As the boundaries between data scientists and MLEs become blurred, knowledge of model deployment will take researchers a long way. In this chapter, you were introduced to a simple workflow for deploying the built model using some open source and easy-to-implement tools. These tools are easy to use and allow you to deploy a web app that can predict TFBS in a quick...