Automated analysis of high‐content microscopy data with deep learning

Abstract Existing computational pipelines for quantitative analysis of high‐content microscopy data rely on traditional machine learning approaches that fail to accurately classify more than a single dataset without substantial tuning and training, requiring extensive analysis. Here, we demonstrate...

Full description

Saved in:
Bibliographic Details
Main Authors: Oren Z Kraus, Ben T Grys, Jimmy Ba, Yolanda Chong, Brendan J Frey, Charles Boone, Brenda J Andrews
Format: Article
Language:English
Published: Springer Nature 2017-04-01
Series:Molecular Systems Biology
Subjects:
Online Access:https://doi.org/10.15252/msb.20177551
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849738362811842560
author Oren Z Kraus
Ben T Grys
Jimmy Ba
Yolanda Chong
Brendan J Frey
Charles Boone
Brenda J Andrews
author_facet Oren Z Kraus
Ben T Grys
Jimmy Ba
Yolanda Chong
Brendan J Frey
Charles Boone
Brenda J Andrews
author_sort Oren Z Kraus
collection DOAJ
description Abstract Existing computational pipelines for quantitative analysis of high‐content microscopy data rely on traditional machine learning approaches that fail to accurately classify more than a single dataset without substantial tuning and training, requiring extensive analysis. Here, we demonstrate that the application of deep learning to biological image data can overcome the pitfalls associated with conventional machine learning classifiers. Using a deep convolutional neural network (DeepLoc) to analyze yeast cell images, we show improved performance over traditional approaches in the automated classification of protein subcellular localization. We also demonstrate the ability of DeepLoc to classify highly divergent image sets, including images of pheromone‐arrested cells with abnormal cellular morphology, as well as images generated in different genetic backgrounds and in different laboratories. We offer an open‐source implementation that enables updating DeepLoc on new microscopy datasets. This study highlights deep learning as an important tool for the expedited analysis of high‐content microscopy data.
format Article
id doaj-art-7d17f04c8f62472f87fec588070afc4e
institution DOAJ
issn 1744-4292
language English
publishDate 2017-04-01
publisher Springer Nature
record_format Article
series Molecular Systems Biology
spelling doaj-art-7d17f04c8f62472f87fec588070afc4e2025-08-20T03:06:36ZengSpringer NatureMolecular Systems Biology1744-42922017-04-0113411510.15252/msb.20177551Automated analysis of high‐content microscopy data with deep learningOren Z Kraus0Ben T Grys1Jimmy Ba2Yolanda Chong3Brendan J Frey4Charles Boone5Brenda J Andrews6Department of Electrical and Computer Engineering, University of TorontoDonnelly Centre for Cellular and Biomolecular Research, University of TorontoDepartment of Electrical and Computer Engineering, University of TorontoCellular Pharmacology, Discovery Sciences, Janssen Pharmaceutical Companies, Johnson & JohnsonDepartment of Electrical and Computer Engineering, University of TorontoDonnelly Centre for Cellular and Biomolecular Research, University of TorontoDonnelly Centre for Cellular and Biomolecular Research, University of TorontoAbstract Existing computational pipelines for quantitative analysis of high‐content microscopy data rely on traditional machine learning approaches that fail to accurately classify more than a single dataset without substantial tuning and training, requiring extensive analysis. Here, we demonstrate that the application of deep learning to biological image data can overcome the pitfalls associated with conventional machine learning classifiers. Using a deep convolutional neural network (DeepLoc) to analyze yeast cell images, we show improved performance over traditional approaches in the automated classification of protein subcellular localization. We also demonstrate the ability of DeepLoc to classify highly divergent image sets, including images of pheromone‐arrested cells with abnormal cellular morphology, as well as images generated in different genetic backgrounds and in different laboratories. We offer an open‐source implementation that enables updating DeepLoc on new microscopy datasets. This study highlights deep learning as an important tool for the expedited analysis of high‐content microscopy data.https://doi.org/10.15252/msb.20177551deep learninghigh‐content screeningimage analysismachine learningSaccharomyces cerevisiae
spellingShingle Oren Z Kraus
Ben T Grys
Jimmy Ba
Yolanda Chong
Brendan J Frey
Charles Boone
Brenda J Andrews
Automated analysis of high‐content microscopy data with deep learning
Molecular Systems Biology
deep learning
high‐content screening
image analysis
machine learning
Saccharomyces cerevisiae
title Automated analysis of high‐content microscopy data with deep learning
title_full Automated analysis of high‐content microscopy data with deep learning
title_fullStr Automated analysis of high‐content microscopy data with deep learning
title_full_unstemmed Automated analysis of high‐content microscopy data with deep learning
title_short Automated analysis of high‐content microscopy data with deep learning
title_sort automated analysis of high content microscopy data with deep learning
topic deep learning
high‐content screening
image analysis
machine learning
Saccharomyces cerevisiae
url https://doi.org/10.15252/msb.20177551
work_keys_str_mv AT orenzkraus automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT bentgrys automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT jimmyba automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT yolandachong automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT brendanjfrey automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT charlesboone automatedanalysisofhighcontentmicroscopydatawithdeeplearning
AT brendajandrews automatedanalysisofhighcontentmicroscopydatawithdeeplearning