Download the full version of the ebook now at ebooksecure.

com

Deep Learning in Bioinformatics: Techniques

and Applications in Practice - eBook PDF

https://ebooksecure.com/download/deep-learning-in-
bioinformatics-techniques-and-applications-in-
practice-ebook-pdf/

Explore and download more ebook at https://ebooksecure.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Artificial Intelligence and Deep Learning in Pathology 1st

Edition Stanley Cohen Md (Editor) - eBook PDF

https://ebooksecure.com/download/artificial-intelligence-and-deep-
learning-in-pathology-ebook-pdf/

ebooksecure.com

(eBook PDF) Bioinformatics and Data Analysis in

Microbiology

https://ebooksecure.com/product/ebook-pdf-bioinformatics-and-data-
analysis-in-microbiology/

ebooksecure.com

(eBook PDF) Data Mining for Business Analytics: Concepts,

Techniques, and Applications in R

https://ebooksecure.com/product/ebook-pdf-data-mining-for-business-
analytics-concepts-techniques-and-applications-in-r/

ebooksecure.com

(eBook PDF) Creating Meaning Through Literature and the

Arts: Arts Integration for Classroom Teachers 5th Edition,

https://ebooksecure.com/product/ebook-pdf-creating-meaning-through-
literature-and-the-arts-arts-integration-for-classroom-teachers-5th-
edition/
ebooksecure.com
Marieb Human Anatomy & Physiology (12th edition) Katja
Hoehn - eBook PDF

https://ebooksecure.com/download/marieb-human-anatomy-physiology-12th-
edition-ebook-pdf/

ebooksecure.com

Introductory Circuit Analysis, 13th Edition (eBook PDF)

https://ebooksecure.com/product/introductory-circuit-analysis-13th-
edition-ebook-pdf/

ebooksecure.com

(eBook PDF) Judgment in Managerial Decision Making 8th by

Bazerman

https://ebooksecure.com/product/ebook-pdf-judgment-in-managerial-
decision-making-8th-by-bazerman/

ebooksecure.com

Pathophysiology 6th edition. Edition Jacquelyn L. Banasik

- eBook PDF

https://ebooksecure.com/download/pathophysiology-ebook-pdf/

ebooksecure.com

(eBook PDF) The Macro Economy Today 14th Edition

https://ebooksecure.com/product/ebook-pdf-the-macro-economy-
today-14th-edition/

ebooksecure.com
Palladacycles: Catalysis and Beyond 1st Edition- eBook
PDF

https://ebooksecure.com/download/palladacycles-catalysis-and-beyond-
ebook-pdf/

ebooksecure.com
Deep Learning in
Bioinformatics
This page intentionally left blank
 Deep Learning in
Bioinformatics
Techniques and Applications in
Practice

Habib Izadkhah
Department of Computer Science
University of Tabriz
Tabriz, Iran
Academic Press is an imprint of Elsevier
125 London Wall, London EC2Y 5AS, United Kingdom
525 B Street, Suite 1650, San Diego, CA 92101, United States
50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom
Copyright © 2022 Elsevier Inc. All rights reserved.
No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and retrieval system, without permission in writing from the
publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our
arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found
at our website: www.elsevier.com/permissions.
This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may
be noted herein).
Notices
Knowledge and best practice in this field are constantly changing. As new research and experience broaden our
understanding, changes in research methods, professional practices, or medical treatment may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any
information, methods, compounds, or experiments described herein. In using such information or methods they should be
mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any
injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

ISBN: 978-0-12-823822-6

For information on all Academic Press publications

visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mara Conner

Acquisitions Editor: Chris Katsaropoulos
Editorial Project Manager: Joshua Mearns
Production Project Manager: Nirmala Arumugam
Designer: Victoria Pearson
Typeset by VTeX
 To my wife Sepideh
and my children Amir Reza and Rose
This page intentionally left blank
Contents

Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xv
CHAPTER 1 Why life science? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Why deep learning? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Contemporary life science is about data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Deep learning and bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 What will you learn? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
CHAPTER 2 A review of machine learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 What is machine learning? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.3 Challenge with machine learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.4 Overfitting and underfitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
2.4.1 Mitigating overfitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4.2 Adjusting parameters using cross-validation . . . . . . . . . . . . . . . . . . . . 15
2.4.3 Cross-validation methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
2.5 Types of machine learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.5.1 Supervised learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.5.2 Unsupervised learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.5.3 Reinforcement learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.6 The math behind deep learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.1 Tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.6.2 Relevant mathematical operations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.6.3 The math behind machine learning: statistics . . . . . . . . . . . . . . . . . . . . 25
2.7 TensorFlow and Keras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.8 Real-world tensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
CHAPTER 3 An introduction of Python ecosystem for deep learning . . . . . . . . . . . . . . . . . 31
3.1 Basic setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2 SciPy (scientific Python) ecosystem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Scikit-learn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4 A quick refresher in Python . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.4.1 Identifier . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.2 Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.3 Data type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.4.4 Control flow statements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.4.5 Data structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.4.6 Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
vii
viii Contents

3.5 NumPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.6 Matplotlib crash course . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.7 Pandas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.8 How to load dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.8.1 Considerations when loading CSV data . . . . . . . . . . . . . . . . . . . . . . . . 46
3.8.2 Pima Indians diabetes dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.8.3 Loading CSV files in NumPy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.8.4 Loading CSV files in Pandas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3.9 Dimensions of your data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
3.10 Correlations between features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.11 Techniques to understand each feature in the dataset . . . . . . . . . . . . . . . . . . . . 53
3.11.1 Histograms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.11.2 Box-and-whisker plots . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.11.3 Correlation matrix plot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
3.12 Prepare your data for deep learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.12.1 Scaling features to a range . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.12.2 Data normalizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
3.12.3 Binarize data (make binary) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.13 Feature selection for machine learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.13.1 Univariate selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.13.2 Recursive feature elimination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.13.3 Principal component analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.13.4 Feature importance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.14 Split dataset into training and testing sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
3.15 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
CHAPTER 4 Basic structure of neural networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
4.2 The neuron . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.3 Layers of neural networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
4.4 How a neural network is trained? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.5 Delta learning rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
4.6 Generalized delta rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.7 Gradient descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.7.1 Stochastic gradient descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80
4.7.2 Batch gradient descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.7.3 Mini-batch gradient descent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
4.8 Example: delta rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.8.1 Implementation of the SGD method . . . . . . . . . . . . . . . . . . . . . . . . . . 86
4.8.2 Implementation of the batch method . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.9 Limitations of single-layer neural networks . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
CHAPTER 5 Training multilayer neural networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
 Contents ix

5.2 Backpropagation algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96

5.3 Momentum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
5.4 Neural network models in keras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
5.5 ‘Hello world!’ of deep learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
5.6 Tuning hyperparameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.7 Data preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.7.1 Vectorization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.7.2 Value normalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
5.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
CHAPTER 6 Classification in bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.1.1 Binary classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
6.1.2 Pima indians onset of diabetes dataset . . . . . . . . . . . . . . . . . . . . . . . . . 115
6.1.3 Label encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
6.2 Multiclass classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
6.2.1 Sigmoid and softmax activation functions . . . . . . . . . . . . . . . . . . . . . . 128
6.2.2 Types of classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
6.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 130
CHAPTER 7 Introduction to deep learning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
7.2 Improving the performance of deep neural networks . . . . . . . . . . . . . . . . . . . 132
7.2.1 Vanishing gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
7.2.2 Overfitting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
7.2.3 Computational load . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150
7.3 Configuring the learning rate in keras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
7.3.1 Adaptive learning rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
7.3.2 Layer weight initializers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
7.4 Imbalanced dataset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
7.5 Breast cancer detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
7.5.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
7.5.2 Introduction and task definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155
7.5.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
7.6 Molecular classification of cancer by gene expression . . . . . . . . . . . . . . . . . . 163
7.6.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.6.2 Introduction and task definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
7.6.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
7.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
CHAPTER 8 Medical image processing: an insight to convolutional neural networks . . . . 175
8.1 Convolutional neural network architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
8.2 Convolution layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
8.3 Pooling layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
8.4 Stride and padding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
8.5 Convolutional layer in keras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Visit https://testbankfan.com
now to explore a rich
collection of testbank or
solution manual and enjoy
exciting offers!
x Contents

8.6 Coronavirus (COVID-19) disease diagnosis . . . . . . . . . . . . . . . . . . . . . . . . . . 184

8.6.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.6.2 Introduction and task definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.6.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
8.6.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
8.7 Predicting breast cancer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8.7.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8.7.2 Introduction and task definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
8.7.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
8.7.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.8 Diabetic retinopathy detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.8.1 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.8.2 Introduction and task definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
8.8.3 Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
8.8.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210
8.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
CHAPTER 9 Popular deep learning image classifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
9.2 LeNet-5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
9.3 AlexNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
9.4 ZFNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
9.5 VGGNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
9.6 GoogLeNet/inception . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
9.7 ResNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
9.8 DenseNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242
9.9 SE-Net . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 245
9.10 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
CHAPTER 10 Electrocardiogram (ECG) arrhythmia classification . . . . . . . . . . . . . . . . . . . . 249
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249
10.2 MIT-BIH arrhythmia database . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
10.3 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
10.4 Data augmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
10.5 Architecture of the CNN model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
10.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
CHAPTER 11 Autoencoders and deep generative models in bioinformatics . . . . . . . . . . . . 261
11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
11.2 Autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
11.2.1 Encoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
11.2.2 Decoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
11.2.3 Distance function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
11.3 Variant types of autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
11.3.1 Undercomplete autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 266
11.3.2 Deep autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
 Contents xi

11.3.3 Convolutional autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

11.3.4 Sparse autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
11.3.5 Denoising autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
11.3.6 Variational autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
11.3.7 Contractive autoencoders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
11.4 An example of denoising autoencoders – bone suppression in chest radiographs 284
11.4.1 Architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
11.5 Implementation of autoencoders for chest X-ray images (pneumonia) . . . . . . . 290
11.5.1 Undercompleted autoencoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
11.5.2 Sparse autoencoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295
11.5.3 Denoising autoencoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 296
11.5.4 Variational autoencoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
11.5.5 Contractive autoencoder . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
11.6 Generative adversarial network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
11.6.1 GAN network architecture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 308
11.6.2 GAN network cost function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
11.6.3 Cost function optimization process in GAN . . . . . . . . . . . . . . . . . . . . . 310
11.6.4 General GAN training process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 310
11.7 Convolutional generative adversarial network . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.7.1 Deconvolution layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.7.2 DCGAN network structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
11.8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
CHAPTER 12 Recurrent neural networks: generating new molecules and proteins sequence
classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
12.2 Types of recurrent neural network . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322
12.3 The problem, short-term memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323
12.4 Bidirectional LSTM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
12.5 Generating new molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 328
12.5.1 Simplified molecular-input line-entry system . . . . . . . . . . . . . . . . . . . 329
12.5.2 A generative model for molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . 330
12.5.3 Generating new SMILES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 336
12.5.4 Analyzing the generative model’s output . . . . . . . . . . . . . . . . . . . . . . . 337
12.6 Protein sequence classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 337
12.6.1 Protein structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
12.6.2 Protein function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
12.6.3 Prediction of protein function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
12.6.4 LSTM with dropout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
12.6.5 LSTM with bidirectional and CNN . . . . . . . . . . . . . . . . . . . . . . . . . . . 344
12.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 346
CHAPTER 13 Application, challenge, and suggestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
13.2 Legendary deep learning architectures, CNN, and RNN . . . . . . . . . . . . . . . . . 347
xii Contents

13.3 Deep learning applications in bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . 348

13.4 Biological networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
13.4.1 Learning tasks on graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
13.4.2 Graph neural networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 352
13.5 Perspectives, limitations, and suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
13.6 DeepChem, a powerful library for bioinformatics . . . . . . . . . . . . . . . . . . . . . . 357
13.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359
Acknowledgments

This book is a product of sincere cooperation of many people. The author would like to thank all those
who contributed in the process of writing and publishing this book. Dr. Masoud Kargar, Dr. Masoud
Aghdasifam, Hamed Babaei, Mahsa Famil, Esmaeil Roohparver, Mehdi Akbari, Mahsa Hashemzadeh,
and Shabnam Farsiani have read the whole draft and made numerous suggestions, improving the pre-
sentation quality of the book; I thank them for all their effort and encouragement.
I wish to express my sincere appreciation to the team at Elsevier, particularly Chris Katsaropoulos,
Senior Acquisitions Editor, for his guidance, comprehensive explanations of the issues, prompt reply
to my e-mails, and, of course, his patience. I would like to also thank Joshua Mearns and Nirmala
Arumugam, for preparing the production process and coordinating the web page, and the production
team. Finally, I thank “the unknown reviewers,” for their great job on exposing what needed to be
restated, clarified, rewritten, and/or complemented.

Habib Izadkhah

xiii
This page intentionally left blank
Exploring the Variety of Random
Documents with Different Content
The Project Gutenberg eBook of In old
Narragansett; romances and realities
This ebook is for the use of anyone anywhere in the United
States and most other parts of the world at no cost and with
almost no restrictions whatsoever. You may copy it, give it away
or re-use it under the terms of the Project Gutenberg License
included with this ebook or online at www.gutenberg.org. If you
are not located in the United States, you will have to check the
laws of the country where you are located before using this
eBook.

Title: In old Narragansett; romances and realities

Author: Alice Morse Earle

Release date: October 13, 2023 [eBook #71873]

Language: English

Original publication: New York: Charles Scribner's Son, 1898

Credits: The Online Distributed Proofreading Team at

https://www.pgdp.net (This file was produced from
images generously made available by The Internet
Archive)

*** START OF THE PROJECT GUTENBERG EBOOK IN OLD

NARRAGANSETT; ROMANCES AND REALITIES ***
 THE IVORY SERIES

Each, 16mo, gilt top, 75 cents

AMOS JUDD. By J. A. Mitchell, Editor of “Life”

IA. A Love Story. By Q. [Arthur T. Quiller-Couch]
THE SUICIDE CLUB. By Robert Louis Stevenson
IRRALIE’S BUSHRANGER. By E. W. Hornung
A MASTER SPIRIT. By Harriet Prescott Spofford
MADAME DELPHINE. By George W. Cable
ONE OF THE VISCONTI. By Eva Wilder Brodhead
A BOOK OF MARTYRS. By Cornelia Atwood Pratt
A BRIDE FROM THE BUSH. By E. W. Hornung
THE MAN WHO WINS. By Robert Herrick
AN INHERITANCE. By Harriet Prescott Spofford
THE OLD GENTLEMAN OF THE BLACK STOCK. By
Thomas Nelson Page
LITERARY LOVE LETTERS AND OTHER STORIES. By
Robert Herrick
A ROMANCE IN TRANSIT. By Francis Lynde
IN OLD NARRAGANSETT. By Alice Morse Earle.
SEVEN MONTHS A PRISONER. By J. V. Hadley.

Other Volumes to be announced

IN OLD NARRAGANSETT

ROMANCES AND REALITIES

BY
ALICE MORSE EARLE

CHARLES SCRIBNER’S SONS

NEW YORK, 1898
 Copyright, 1898, by
Charles Scribner’s Sons

TROW DIRECTORY
PRINTING AND BOOKBINDING COMPANY
NEW YORK
 FOREWORD
Some of these stories of old Narragansett are familiar fireside
tales to those who have lived in that picturesque land; some are but
vague traditions, others summer dreams; a few are family chronicles;
still others are outlined in that interesting memoir, Thomas R.
Hazard’s “Recollections of the Olden Times,” or in Updike’s
“Narragansett Church.” Old Narragansett was, properly, all the lands
occupied by the Narragansett Indians at the coming of the English.
Narragansett is now, popularly, the coast sweep of the western shore
of Narragansett Bay from Wickford to Point Judith. In 1685
Narragansett was made a separate government apart from Rhode
Island, and was called the Kings Province. When reunited with
Rhode Island this was changed to King’s County. For many years,
and by some old people to-day, it is called the South County, but its
legal name is Washington County, which was given it in 1781;
Washington being a more agreeable and tolerable name at that date
to loyal Americans than King’s. Narragansett was owned by a
comparatively small number of persons, and estates were large, one
family owned a tract nine miles long and three wide. Thomas
Stanton had a “lordship” four and a half miles long and two wide.
Colonel Champlin owned two thousand acres, Thomas Hazard
twelve thousand acres. Farms of five, six, even ten miles square
existed.
Thus the conditions of life in colonial Narragansett were widely
different from those of other New England colonies. The
establishment of and adherence to the Church of England, and the
universal prevalence of African slavery, evolved a social life
resembling that of the Virginian plantation rather than of the Puritan
farm. It was a community of many superstitions, to which the folk-
customs of the feast-days of the English Church, the evil
communications of witch-seeking Puritan neighbors, the voodooism
of the negro slaves, the pow wows of the native red men, all added a
share and infinite variety. It was a plantation of wealth, of vast flocks
and herds, of productive soil, of great crops, of generous living; all
these are vanished from the life there to-day, but still the fields are
smiling and the lakes and the bay are blue and beautiful as of yore;
and a second prosperity is dawning in the old Kings Province in the
universal establishment therein of happy summer-homes.
In memory of many perfect days spent on Narragansett roads and
lanes, of days in Narragansett woods or on the shore, these pages
have been written.
Alice Morse Earle.
Wickford, Rhode Island,
Midsummer Eve, 1897.
 CONTENTS
PAGE
A Narragansett Elopement 1
Narragansett Weavers 23
Where Three Towns Meet 51
Tuggie Bannocks’s Moonack 63
A Black Politician 77
The Witch Sheep 103
The Crusoes of the Noon-House 121
The Doctor’s Pie-Plates 139
My Delft Apothecary Jars 151
The Dancing Turkey 169
Cuddymonk’s Ghost 181
 A NARRAGANSETT ELOPEMENT
Four miles north of Narragansett Pier lies the old South Ferry, from
whence for over a century ran ferry-boats to a landing on Conanicut
Island. About a mile farther north there stands on Boston Neck an
ancient willow-shaded, gambrel-roofed, weather-beaten house which
in the latter part of the eighteenth century was the scene of a sadly
romantic event. It was built by Rowland Robinson in the first half of
the century—in 1746—and was originally one hundred and ten feet
long, as the stone foundations still show. The kitchen and negro
quarters have been demolished, and the present structure has a
front of sixty feet. The rooms within are models of the simple style of
architecture of that day. The staircase is specially beautiful with its
gracefully turned balusters and curious drop ornaments, and its
deep-worn steps of bass-wood. The walls of all the rooms are
wainscoted in a substantial manner, and the fireplaces are
ornamented with blue and white Dutch tiles. The heavy timbers and
rafters—all cut on the place—have not sagged an inch with the
weight of years. Over the fireplace in the dining-room is a panel
bearing a smoke-darkened painting which represents a deer-hunt
that occurred on the Robinson place while the house was being built.
The riders in this picture appear to be standing in their stirrups
instead of sitting on their saddles. The great attic in which the slaves
are said to have slept contains now a picturesque litter of old sea-
chests, spinning-wheels, clock-reels, wool-cards, flax-brakes, yarn-
winders, saddles, and pillions; and in the beams of the roof are great
iron hooks to which—it is whispered—the slaves of olden times were
tied when they received their floggings. They are with much more
probability the loom-hooks which were used by weavers when
weaving cloth on an old hand-loom. The handsome great west room
is known as the Lafayette Chamber, it having been occupied for
some weeks by the Marquis de Lafayette during the Revolutionary
War; and on panes of glass, still whole after a century’s use, are the
names of French officers, scratched on with the writers’ diamond
rings.
The house abounds in cupboards—tall, narrow cupboards high up
over the chimney, low, broad cupboards under the window-seats,
medicine cupboards and pot cupboards, triangular corner
cupboards, and, in the parlor, one beautifully proportioned apse-
shaped china-cupboard which is ornamented with carved “sunbursts”
and scalloped and serrated shelves, and is closed with glass doors
to show the treasures and beauties within. But in “Unfortunate
Hannah’s” chamber is the most famous cupboard of all, for in that
narrow and shallow retreat a beautiful daughter of Rowland
Robinson hid her lover when she heard the approaching footsteps of
her irascible father on the staircase leading to her room.
Rowland Robinson was a typical Narragansett planter—wealthy,
proud, and imperious. Tall and portly, ruddy of face, he showed in his
dress and carriage his great wealth and high position. A coat of fine
dark cloth or velvet with silver buttons was worn over a long yellow
waistcoat with great pockets and flaps; violet or brown velvet knee-
breeches with handsome top-boots, or silk stockings with buckled
shoes; lace-frilled shirts; a great beaver cocked hat looped up with
cords over his powdered hair—this attire gave him a comely and
elegant presence. His character may be given in a few words by
quoting the wife of Hon. William Hunter, minister to Brazil. She wrote
in her diary sixty years ago her personal recollection of him. “He was
of violent passions, which was characteristic of the Robinsons, but of
benevolent, noble nature.” Many stories are told of his impetuous
generosity and kindly impulsiveness, none being more characteristic
than his action when his first cargo of slaves came from the Guinea
coast. Slave-dealing was such a universal practice at that date
among wealthy residents of Narragansett and Newport that it was a
commonplace business enterprise for Rowland Robinson, when he
was building his new house, to send a ship to Africa for a cargo of
negroes, intending to keep the most promising ones for his own
household and farm servants, and to sell the remainder. But when
the ship landed at South Ferry, and the forlorn, wretched, feeble men
and women disembarked, he burst into tears and vowed that not one
should be sold. He kept them all in his own household, where they
were always kindly treated. He never again sent a vessel to Africa to
engage in the slave-trade, though one negro of royal birth—Queen
Abigail—was so happy in her Narragansett home, that with Rowland
Robinson’s consent and his liberal assistance she returned to her
home in Africa, found her son—the negro prince—and brought him
to America, where he became Mr. Robinson’s faithful body-servant.
The wealthy planter had other sources of income than slave-
trading. He owned great ships that engaged in general commerce.
He had an immense dairy and made fine Rhode Island cheese from
the milk of his beautiful “blanket-cows.” It was his ambition to have
one hundred of these lovely black-and-white animals, but it is a
matter of tradition that, while he could keep ninety-nine readily
enough, when he bought or raised the hundredth cow, one of the
ninety-nine sickened and died, or was lost through accident, and
thus the number still fell short. Great quantities of grain and hay did
he also raise on his fertile farm; and besides the grain and cheese
that he shipped to the West Indies he also sold to the wealthy
colonists many Narragansett pacers—swift horses of the first
distinctively American breed. These pacers all came from one sire,
“Old Snip,” who it is said was of Andalusian birth and was found
swimming in the ocean off the coast of Africa, was hauled on board a
trading-ship and was carried to Narragansett, where he was allowed
to run wild on the Point Judith tract. These sure-footed pacers had a
peculiar gait; they did not sway their rider from side to side, nor jolt
him up and down, but permitted him to sit quietly, and thus endure
without fatigue a long journey. In those carriageless days, when
nearly all travel was by saddle and pillion, the broad-backed, easy-
going Narragansett pacers were in such demand that they brought
high prices and proved a good source of income.
Three children were born to the builder of this beautiful colonial
home: William Robinson, who died in Newport in 1804, in a house on
the corner of Broadway and Mann Avenue, and two daughters, Mary
and Hannah. Gay festivities had these young people in the
hospitable great house, especially when a demure young Quaker
cousin was sent to them to live for awhile in order to break up a
romantic love-affair of hers with a young French officer. Count
Rochambeau was a guest at her father’s house, and too many
opportunities for love-making were found when the young
Frenchman came to report to his commanding officer.
Gayest and loveliest of all the beauties throughout Narragansett
was fair Hannah Robinson—Unfortunate Hannah. Much testimony of
her extraordinary beauty has descended to us, one story being of
her meeting with Crazy Harry Babcock, that reckless dare-devil of a
soldier whose feats of valor by land and sea were known all over
Europe as well as in America. This extraordinary man, during a visit
to England, was invited to the palace and introduced to the royal
family. When the queen extended her hand to him to be kissed, he
sprang briskly from his knees, exclaiming: “May it please your
majesty, in my country, when we salute a beautiful woman we kiss
her lips, not her hand,” and with the words he seized the astonished
queen by the shoulders and impressed on her lips a rousing smack.
Upon his return to America he went to Narragansett for the avowed
purpose of “seeing the prettiest woman in Rhode Island.” As he
entered the parlor of Rowland Robinson’s house fair Hannah rose to
meet him, and the crazy colonel, as she extended her hand to greet
him, dropped on his knee with a look of intense admiration, saying,
in the stilted words of the times: “Pray permit one who has kissed
unrebuked the lips of the proudest queen on earth to press for a
moment the hand of an angel from heaven!”
The great wealth and luxurious manner of living of the opulent
Narragansett planters was shown in no way more plainly than in the
manner in which they educated their children. They spared no pains
nor expense to obtain the best masters and teachers. Rowland
Robinson sent his daughter to Newport to receive instruction from
Madame Osborne, whose fame as a teacher was known throughout
America, and whose “Memoirs” form the dullest book in the English
language. At this school Hannah met the handsome lover who was
to have such an influence over her life. Pierre Simond, or Peter
Simons as was most unromantically Anglicized his name, was a
scion of a French Huguenot family, who taught music and French in
Madame Osborne’s school. From the moment the young couple met
they were lovers. Both knew, however, how hopeless it was to think
of obtaining Mr. Robinson’s consent to a marriage which would
appear to him so unequal; they therefore kept their love a secret.
As the time approached for Hannah’s return to her home in
Narragansett, the lovers were in despair at the thought of separation,
for they knew their unhappiness could not be mitigated even by the
exchange of love-letters. At this juncture the young music-teacher
managed to obtain a position as private tutor in the family of Colonel
Gardiner, who lived only two miles from Hannah’s home and who
was her uncle.
It can easily be divined that when once in Narragansett the happy
lovers found many opportunities of meeting, which were frequently
brought about by the romantic and easy-going colonel, and were not
hindered by Hannah’s mother when she discovered her daughter’s
love-affair. Though Mrs. Robinson would not give her approval she
tacitly gave her aid by helping to conceal the lovers’ meetings from
Rowland Robinson; and it was with her knowledge that the lover
came to Hannah’s chamber, where he often had to be concealed in
the friendly cupboard.
When Peter Simons could not enter the Robinson house he stood
by his true-love’s window under a great lilac-bush, which is still
growing, sturdy and unbroken under the weight of a century of years.
In the concealing shadow of the lilac-bush words of love might be
whispered to the fair girl who leaned from the window, or letters
might be exchanged with comparative safety.
But true love ran no smoother in the eighteenth century than in the
nineteenth, and when one night a fair hand dropped a tender billet
into the gloom of the lilac-bush, old Rowland Robinson chanced to
open the door of his house and he saw the white messenger
descend. Speechless with suspicion and rage he rushed to the lilac-
bush and thrust his buckthorn stick into it with vigorous blows until a
man ran out into the darkness, whom the irate father in the second’s
glimpse recognized as the “wretched French dancing-master” who
taught his nephews.
 The horrified and disgusted anger of Rowland Robinson and the
scene that ensued within doors can well be imagined; little peace or
happiness was there for Hannah after her father’s discovery. Updike,
in his “History of the Narragansett Church,” says of her life at this
time: “If she walked, her movements were watched; if she rode, a
servant was ordered to be in constant attendance; if a visit was
contemplated, her father immediately suspected it was only a pretext
for an arranged interview; and even after departure, if the most
trifling circumstance gave color to suspicion, he would immediately
pursue and compel her to return. In one instance she left home to
visit her aunt in New London; her father soon afterward discovered
from his windows a vessel leaving Newport and taking a course
toward the same place. Although the vessel and the persons on
board were wholly unknown to him, his jealousies were immediately
aroused. Conjecturing it was Mr. Simons intending to fulfil an
arrangement previously made, he hastened to New London, arrived
a few hours only after his daughter, and insisted on her instant
return. No persuasions or argument could induce him to change his
determination, and she was compelled to return with him.”
Though Rowland Robinson was firm in his determination and
constant in his action to prevent the lovers from meeting, Hannah—
the true daughter of her father—was equally determined not to give
up her sweetheart; and as the Narragansett neighbors, like the rest
of the world, “dearly loved a lover,” they gladly assisted the romance
by exchanging letters and arranging meetings for the lovers. Months
of harassing suspicions and angry words at home, and frightened
meetings with her lover away from home, told so upon Hannah’s
health that her mother finally permitted to be carried into execution a
long-planned scheme of elopement. It was finally arranged through
the agency and assistance of a young friend of Hannah’s—Miss
Belden—and the ever sentimental colonel-uncle.
Invitations for a great ball had been sent out all over Narragansett,
and to many in Boston, Providence, and Newport. It was to be given
by Mrs. Updike, Hannah’s aunt. She lived eight miles north of
Rowland Robinson’s home, in the old historic house which is still
standing and is now known as Cocumcussuc. A portion of it was the
first house or fort built by the English in Narragansett in the year
1636. Though Hannah’s father was unwilling to allow his daughter
out of his sight, he at last consented that both Hannah and Mary
should go to their aunt’s ball. They set out on horseback,
accompanied by faithful Prince, the son of Queen Abigail, and were
met, as had been arranged, in the thick woods on the top of Ridge
Hill, by Mr. Simons with a closed carriage. Into this conveyance
Hannah entered with her lover, in spite of her sister’s tears and
Prince’s frantic appeals, and rode off to Providence, where the
eloping couple were married.
When the news of Hannah’s disobedience came to the knowledge
of Rowland Robinson, his rage and disgust knew no bounds. He
forbade his family ever to communicate with Hannah again; and
knowing well that she must have been assisted in carrying out her
plans to elope, he offered a large reward to anyone who would make
known to him the names of the persons who had aided her escape.
It would seem that the fair bride should be called Fortunate
Hannah, since she managed to evade her father’s vigilance and wed
her ardent French lover, but alas! Peter Simons, like many another
hero of an elopement, did not prove worthy of the great sacrifice.
Disappointed through the implacable anger of Rowland Robinson in
the hope of obtaining any of his wealth, the unprincipled husband
soon neglected his lovely wife and at last deserted her for days and
weeks. Broken-hearted, alone, and poor, the unfortunate girl began
to fail rapidly in health, and spent many weary, lonely days in her
wretched home in Providence, having for her only companion her
dog Marcus, that had been secretly sent to her by her mother from
her Narragansett home.
In the meantime her sister, Mary Robinson, had died of
consumption; and her mother, worn out by grief, had completely
failed in health. Her father, though outwardly stern and unforgiving,
was evidently exceedingly unhappy at the alarming news of his
daughter’s state of health; and at last, of his own accord, sent to live
with her and care for her the negro maid who had attended her in her
happy girlhood. He also conveyed to her the message that she might
come home and would be warmly welcomed, provided she would
reveal to him the names of those who assisted in her elopement. Her
compliance with this condition was, he said, absolutely imperative.
On receiving this message Hannah wrote in answer, with trembling
hand, a most affectionate letter, stating firmly that the sentiments of
honor which he himself had both taught and transmitted to her
forbade her betraying the confidence of those who had aided her
and offended him. Mr. Robinson eagerly opened the letter, but his
face changed when he read her decision, and he tossed the sheet to
her mother with the contemptuous remark, “Then let the foolish thing
die where she is!”
As weeks passed the accounts of Hannah’s health grew more
alarming still, and it was evident that a fierce struggle between love
and pride was taking place in the unhappy father’s breast; one day
he rose suddenly from the dinner-table, jumped upon his horse, and
saying to his wife that he should be away from home for a day or
two, started on the thirty-five-mile ride to Providence. He remained
overnight at the Updike farm and reached his daughter’s house in
Providence at noon. Without dismounting he rapped on the door with
his riding-whip. Full of joy at the sight of her old master and at the
thought of the happy reconciliation, the negro maid hastened to the
door with the entreaty that the welcome visitor would come at once
to the poor invalid’s chamber. “Ask your mistress,” said Rowland
Robinson, “whether she is now ready to comply with her father’s
request to know the names of her fellow-conspirators, and say that if
she is, he will come in, but on no other conditions.” Poor Hannah,
torn with a thousand emotions, still clung to her decision not to
betray her friends, and her father, without another word, rode away
to the Updike farm. For several weeks the stubborn and unhappy
father, unable to live without news of his sick daughter, rode at
intervals of two or three days from Narragansett to Providence,
knocked at Hannah’s door, asked for her health, and left without
another word.
At last, her friends who had helped in her elopement, hearing of
her father’s firm decision, which barred all reconciliation, insisted
upon her revealing to him their names and the true story; and when
Rowland Robinson next rode up to his daughter’s door he received
the welcome message that she would see him and tell him all. When
he entered that barren chamber all thought of discovering her closely
guarded secret fled at once from his thoughts as he gazed at the
wasted form of the once beautiful girl. He knelt by her bedside and
wept aloud in anguish and remorse. As soon as he recovered his
composure he at once rode to his home, from whence he
despatched to Providence in a fast-sailing sloop four of his strongest
and trustiest negro men, and a hand-litter for the sick, which was, at
that time of rough roads and few carriages, an indispensable article
in every well-appointed Narragansett household. Dusty, travel-
stained, and tired, without waiting for a night’s rest he at once
jumped upon a fresh horse and, attended by Prince, who was
mounted and led a horse for Hannah’s maid, poor Rowland
Robinson started for the last time to ride to his sick daughter’s door.
Upon a lovely morning in June, the four strong negroes, bearing
the litter upon which lay the sick girl, with her father and faithful
Prince riding on either side, slowly wended their way to poor
Hannah’s early home. Those who know the beauty of sunny
Narragansett in early June, when the roads are everywhere
overhung with the graceful, sweet-scented blossoms of slender
locust-trees, when the roadsides are one luxuriant, blooming garden
of lovely wild flowers, and the fields are sweet with rich clover, can
feel the strong and painful contrast which the sad figure of the dying
girl must have formed to the glowing life around.
When the spot was reached on Ridge Hill where Hannah had seen
for the last time her sister Mary, Prince saw that she covered her
face with her hands and cried. One other pathetic incident is told by
“Shepherd Tom” of the homeward journey. Though on every side lay
a glory of spring flowers, poor Hannah, with thoughts that no one can
fathom, asked her father to pick for her and lay on her breast a
withered sprig of the pale blossom called life-everlasting, which had
bloomed and died the year before.
At last the painful journey was ended; of the sad meeting between
mother and daughter, and of the sorrowful faces of the faithful
servants, it is needless to write in detail.
 That night a whip-poor-will—the bird believed throughout
Narragansett to be the harbinger of death—perched on the lilac-bush
under the window of the chamber where once again slept
Unfortunate Hannah; and throughout the long dark hours sounded
gloomily in the father’s ears the sad, ominous cry of “Whip-poor-will!
Whip-poor-will!” The following day poor Hannah died.
Again did four strong men bear on their shoulders the form of the
once beautiful girl, as they passed under the branches of the sweet-
scented lilac to the grave near the old house where still is shown the
headstone that marks the last resting-place of Unfortunate Hannah
Robinson.