## Problem description This directory contains data to train a model to predict the function of protein domains, based on the PFam dataset. Domains are functional sub-parts of proteins; much like images in ImageNet are pre segmented to contain exactly one object class, this data is presegmented to contain exactly and only one domain. The purpose of the dataset is to repose the PFam seed dataset as a multiclass classification machine learning task. The task is: given the amino acid sequence of the protein domain, predict which class it belongs to. There are about 1 million training examples, and 18,000 output classes. ## Data structure This data is more completely described by the publication "Can Deep Learning Classify the Protein Universe