Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments

Takashi Kamatani, Koichi Fukunaga, Kaede Miyata, Yoshitaka Shirasaki, Junji Tanaka, Rie Baba, Masako Matsusaka, Naoyuki Kamatani, Kazuyo Moro, Tomoko Betsuyaku, Sotaro Uemura

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

For single-cell experiments, it is important to accurately count the number of viable cells in a nanoliter well. We used a deep learning-based convolutional neural network (CNN) on a large amount of digital data obtained as microscopic images. The training set consisted of 103 019 samples, each representing a microscopic grayscale image. After extensive training, the CNN was able to classify the samples into four categories, i.e., 0, 1, 2, and more than 2 cells per well, with an accuracy of 98.3% when compared to determination by two trained technicians. By analyzing the samples for which judgments were discordant, we found that the judgment by technicians was relatively correct although cell counting was often difficult by the images of discordant samples. Based on the results, the system was further enhanced by introducing a new algorithm in which the highest outputs from CNN were used, increasing the accuracy to higher than 99%. Our system was able to classify the data even from wells with a different shape. No other tested machine learning algorithm showed a performance higher than that of our system. The presented CNN system is expected to be useful for various single-cell experiments, and for high-Throughput and high-content screening.

Original languageEnglish
Article number16831
JournalScientific Reports
Volume7
Issue number1
DOIs
Publication statusPublished - 2017 Dec 1

Fingerprint

Learning algorithms
Neural networks
Experiments
Learning systems
Screening
Throughput
Deep learning

ASJC Scopus subject areas

  • General

Cite this

Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments. / Kamatani, Takashi; Fukunaga, Koichi; Miyata, Kaede; Shirasaki, Yoshitaka; Tanaka, Junji; Baba, Rie; Matsusaka, Masako; Kamatani, Naoyuki; Moro, Kazuyo; Betsuyaku, Tomoko; Uemura, Sotaro.

In: Scientific Reports, Vol. 7, No. 1, 16831, 01.12.2017.

Research output: Contribution to journalArticle

Kamatani, T, Fukunaga, K, Miyata, K, Shirasaki, Y, Tanaka, J, Baba, R, Matsusaka, M, Kamatani, N, Moro, K, Betsuyaku, T & Uemura, S 2017, 'Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments', Scientific Reports, vol. 7, no. 1, 16831. https://doi.org/10.1038/s41598-017-17012-x
Kamatani, Takashi ; Fukunaga, Koichi ; Miyata, Kaede ; Shirasaki, Yoshitaka ; Tanaka, Junji ; Baba, Rie ; Matsusaka, Masako ; Kamatani, Naoyuki ; Moro, Kazuyo ; Betsuyaku, Tomoko ; Uemura, Sotaro. / Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments. In: Scientific Reports. 2017 ; Vol. 7, No. 1.
@article{5011ba27aee143c0a1d4afe96de2cfe0,
title = "Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments",
abstract = "For single-cell experiments, it is important to accurately count the number of viable cells in a nanoliter well. We used a deep learning-based convolutional neural network (CNN) on a large amount of digital data obtained as microscopic images. The training set consisted of 103 019 samples, each representing a microscopic grayscale image. After extensive training, the CNN was able to classify the samples into four categories, i.e., 0, 1, 2, and more than 2 cells per well, with an accuracy of 98.3{\%} when compared to determination by two trained technicians. By analyzing the samples for which judgments were discordant, we found that the judgment by technicians was relatively correct although cell counting was often difficult by the images of discordant samples. Based on the results, the system was further enhanced by introducing a new algorithm in which the highest outputs from CNN were used, increasing the accuracy to higher than 99{\%}. Our system was able to classify the data even from wells with a different shape. No other tested machine learning algorithm showed a performance higher than that of our system. The presented CNN system is expected to be useful for various single-cell experiments, and for high-Throughput and high-content screening.",
author = "Takashi Kamatani and Koichi Fukunaga and Kaede Miyata and Yoshitaka Shirasaki and Junji Tanaka and Rie Baba and Masako Matsusaka and Naoyuki Kamatani and Kazuyo Moro and Tomoko Betsuyaku and Sotaro Uemura",
year = "2017",
month = "12",
day = "1",
doi = "10.1038/s41598-017-17012-x",
language = "English",
volume = "7",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
number = "1",

}

TY - JOUR

T1 - Construction of a system using a deep learning algorithm to count cell numbers in nanoliter wells for viable single-cell experiments

AU - Kamatani, Takashi

AU - Fukunaga, Koichi

AU - Miyata, Kaede

AU - Shirasaki, Yoshitaka

AU - Tanaka, Junji

AU - Baba, Rie

AU - Matsusaka, Masako

AU - Kamatani, Naoyuki

AU - Moro, Kazuyo

AU - Betsuyaku, Tomoko

AU - Uemura, Sotaro

PY - 2017/12/1

Y1 - 2017/12/1

N2 - For single-cell experiments, it is important to accurately count the number of viable cells in a nanoliter well. We used a deep learning-based convolutional neural network (CNN) on a large amount of digital data obtained as microscopic images. The training set consisted of 103 019 samples, each representing a microscopic grayscale image. After extensive training, the CNN was able to classify the samples into four categories, i.e., 0, 1, 2, and more than 2 cells per well, with an accuracy of 98.3% when compared to determination by two trained technicians. By analyzing the samples for which judgments were discordant, we found that the judgment by technicians was relatively correct although cell counting was often difficult by the images of discordant samples. Based on the results, the system was further enhanced by introducing a new algorithm in which the highest outputs from CNN were used, increasing the accuracy to higher than 99%. Our system was able to classify the data even from wells with a different shape. No other tested machine learning algorithm showed a performance higher than that of our system. The presented CNN system is expected to be useful for various single-cell experiments, and for high-Throughput and high-content screening.

AB - For single-cell experiments, it is important to accurately count the number of viable cells in a nanoliter well. We used a deep learning-based convolutional neural network (CNN) on a large amount of digital data obtained as microscopic images. The training set consisted of 103 019 samples, each representing a microscopic grayscale image. After extensive training, the CNN was able to classify the samples into four categories, i.e., 0, 1, 2, and more than 2 cells per well, with an accuracy of 98.3% when compared to determination by two trained technicians. By analyzing the samples for which judgments were discordant, we found that the judgment by technicians was relatively correct although cell counting was often difficult by the images of discordant samples. Based on the results, the system was further enhanced by introducing a new algorithm in which the highest outputs from CNN were used, increasing the accuracy to higher than 99%. Our system was able to classify the data even from wells with a different shape. No other tested machine learning algorithm showed a performance higher than that of our system. The presented CNN system is expected to be useful for various single-cell experiments, and for high-Throughput and high-content screening.

UR - http://www.scopus.com/inward/record.url?scp=85037138730&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85037138730&partnerID=8YFLogxK

U2 - 10.1038/s41598-017-17012-x

DO - 10.1038/s41598-017-17012-x

M3 - Article

VL - 7

JO - Scientific Reports

JF - Scientific Reports

SN - 2045-2322

IS - 1

M1 - 16831

ER -