Merge pull request #1996 from deepchem/examples_1

# DeepChem Example Suite

This directory contains the DeepChem example suite. There are a large number of

examples which break into a few broad categories:

- API Examples: These examples show how to do little things with DeepChem's API

  that you might not have realized were possible.

- Case Study Examples: These show how to analyze interesting datasets with DeepChem.

- Tutorial Notebooks: These IPython notebooks provide walkthroughs of using

  DeepChem on interesting problems in practice.

# BACE Dataset Examples

The BACE dataset is from the following paper:

Subramanian, Govindan, et al. "Computational modeling of β-secretase 1 (BACE-1) inhibitors using ligand based approaches." Journal of chemical information and modeling 56.10 (2016): 1936-1949.

This study considers a small dataset of 205 compounds datasets

which are used to train a model which is evaluated on a larger

external validation set of 1273 compounds.

The file `bace_datasets.py` loads the data as used in the

original paper. `bace_rf.py` demonstrates training a random

forest against this dataset.

# Clintox dataset models

The Clintox dataset is a collection of "clinical toxicity" datasets that compares drugs approved by the FDA and drugs that have failed clinical trials for toxicity reasons. It contains two classification tasks for 1491 compounds:

1) Clinical trial toxicity/non-toxicity

2) FDA approval status

In this example, we construct fully connected deep networks and

graph convolutional models for the task of predicting clinical

toxicity/FDA approval status from molecular structure.

Script that trains multitask models on clintox dataset.

@author Caleb Geniesse

"""

from __future__ import print_function

from __future__ import division

from __future__ import unicode_literals

import numpy as np

import deepchem as dc

from deepchem.molnet import load_clintox