Wide and Deep Learning for Automatic Cell Type Identification

AbstractCell type classification is an important problem in cancer research, especially with the advent of single cell technologies. Correctly identifying cells within the tumor microenvironment can provide oncologists with a snapshot of how a patient’s immune system is reacting to the tumor. Wide deep learning (WDL) is an approach to construct a cell-classification prediction model that can learn patterns within high-dimensional data (deep) and ensure that biologically relevant features (wide) remain in the final model. In this paper, we demonstrate that the use of regularization can prevent overfitting and adding a wide component to a neural network can result in a model with better predictive performance. In particular, we observed that a combination of dropout and ℓ2 regularization can lead to a validation loss function that does not depend on the number of training iterations and does not experience a significant decrease in prediction accuracy compared to models with ℓ1, dropout, or no regularization. Additionally, we show WDL can have superior classification accuracy when the training and testing of a model is completed data on that arise from the same cancer type, but from different platforms. More specifically, WDL compared to traditional deep learning models can substantially increase the overall cell type prediction accuracy (41 to 90%) and T-cell sub-types (CD4: 0 to 76%, and CD8: 61 to 96%) when the models were trained using melanoma data obtained from the 10X platform and tested on basal cell carcinoma data obtained using SMART-seq.

Download Full-text

Implications of Epigenetic Variability within a Cell Population for “Cell Type” Classification

Frontiers in Behavioral Neuroscience ◽

10.3389/fnbeh.2015.00342 ◽

2015 ◽

Vol 9 ◽

Cited By ~ 8

Author(s):

Inna Tabansky ◽

Joel N. H. Stern ◽

Donald W. Pfaff

Keyword(s):

Cell Population ◽

Cell Type ◽

Epigenetic Variability ◽

A Cell ◽

Type Classification

Download Full-text

Cell type classification and unsupervised morphological phenotype identification from low-res images with deep learning

10.1101/533216 ◽

2019 ◽

Author(s):

Kai Yao ◽

Nash D. Rochman ◽

Sean X. Sun

Keyword(s):

Single Cell ◽

Cell Density ◽

Semantic Segmentation ◽

Cell Interaction ◽

Support Network ◽

Cell Type ◽

A Cell ◽

Convnet Feature ◽

Expert Classification ◽

Type Classification

AbstractConvolutional neural networks (ConvNets) have been used for both classification and semantic segmentation of cellular images. Here we establish a method for cell type classification utilizing images taken on a benchtop microscope directly from cell culture flasks eliminating the need for a dedicated imaging platform. Significant flask-to-flask heterogeneity was discovered and overcome to support network generalization to novel data. Cell density was found to be a prominent source of heterogeneity even within the single-cell regime indicating the presence of morphological effects due to diffusion-mediated cell-cell interaction. Expert classification was poor for single-cell images and excellent for multi-cell images suggesting experts rely on the identification of characteristic phenotypes within subsets of each population and not ubiquitous identifiers. Finally we introduce Self-Label Clustering, an unsupervised clustering method relying on ConvNet feature extraction able to identify distinct morphological phenotypes within a cell type, some of which are observed to be cell density dependent.Author summaryK.Y., N.D.R., and S.X.S. designed experiments and computational analysis. K.Y. performed experiments and ConvNets design/training, K.Y., N.D.R and S.X.S wrote the paper.

Download Full-text

Frontal Cortex Neuron Type Classification with Deep Learning and Recurrence Plot

Traitement du signal ◽

10.18280/ts.380327 ◽

2021 ◽

Vol 38 (3) ◽

pp. 807-819

Author(s):

Fatma Özcan ◽

Ahmet Alkan

Keyword(s):

Deep Learning ◽

Frontal Cortex ◽

Recurrence Plot ◽

Image Features ◽

Spike Timing ◽

Nerve Cells ◽

Network Activity ◽

Cell Type ◽

Spike Sequences ◽

Type Classification

One of the goals of neural decoding in neuroscience is to create Brain-Computer Interfaces (BCI) that use nerve signals. In this context, we are interested in the activity of nerve cells. It is possible to classify nerve cells as excitatory or inhibitors by evaluating individual extra-cellular measurements taken from the frontal cortex of rats. Classification of neurons with only spike timing values has not been studied before, with deep learning, without knowing all of the wave properties and the intercellular interactions. In this study, inter-spike interval values of individual neuronal spike sequences were converted into recurrence plot images to analyze as point processing, image features were extracted using the pre-trained AlexNet with CNN deep learning method, and frontal cortex nerve cell type classification was made. Kernel classification, SVM, Naive Bayes, Ensemble, decision trees classification methods were used. The accuracy, sensitivity and specificity evaluate the proposed methods. A success of more than 81% has been achieved. Thus, the cell type is defined automatically. It has been observed that the ISI properties of spike trains can carry out information on cell type and thus neural network activity. Under these circumstances, these values are significant and important for neuroscientists.

Download Full-text

DeepBCR: Deep learning framework for cancer-type classification and binding affinity estimation using B cell receptor repertoires

10.1101/731158 ◽

2019 ◽

Author(s):

Xihao Hu ◽

X Shirley Liu

Keyword(s):

Deep Learning ◽

B Cell ◽

Binding Affinity ◽

Cell Receptor ◽

B Cell Receptors ◽

Cancer Type ◽

Binding Affinities ◽

Learning Framework ◽

Kd Values ◽

Type Classification

AbstractWe developed a deep learning framework to model the binding specificity of B-cell receptors (BCRs). The DeepBCR framework can predict the cancer type from a repertoire of BCRs and estimate the binding affinity of a single BCR. We designed a peptide encoding network that includes an amino acid encoding layer, k-mer motif layer, and immunoglobulin isotype layer, and used transfer learning to reduce parameters and over-fitting. When we applied the framework to evaluate the three commercial anti-PD1 drugs (Opdivo, Keytruda, and Libtayo), the predicted binding affinities correlate with the real affinities measured in kD values. This validates the prediction and indicates that we can use the framework to select strong antigen-specific binders.

Download Full-text

Predictive advantage of a cell type classification for pulmonary adenocarcinoma coupled with data for p53, K-ras and EGFR alterations

Cancer Science ◽

10.1111/j.1349-7006.2010.01585.x ◽

2010 ◽

Vol 101 (7) ◽

pp. 1745-1753 ◽

Cited By ~ 3

Author(s):

Akira Okada ◽

Takuo Shimmyo ◽

Takehisa Hashimoto ◽

Yasuhito Kobayashi ◽

Yohei Miyagi ◽

...

Keyword(s):

Pulmonary Adenocarcinoma ◽

Cell Type ◽

A Cell ◽

Type Classification

Download Full-text

Single-cell entropy to quantify the cellular transcriptome from single-cell RNA-seq data

10.1101/678557 ◽

2019 ◽

Author(s):

Jingxin Liu ◽

You Song ◽

Jinzhi Lei

Keyword(s):

Single Cell ◽

Gaussian Mixture ◽

Rna Seq ◽

Cell Type ◽

Transcriptome Profile ◽

Biological Interpretation ◽

A Cell ◽

Transcriptional State ◽

Type Classification ◽

Cellular Transcriptome

We present the use of single-cell entropy (scEntropy) to measure the order of the cellular transcriptome profile from single-cell RNA-seq data, which leads to a method of unsupervised cell type classification through scEntropy followed by the Gaussian mixture model (scEGMM). scEntropy is straightforward in defining an intrinsic transcriptional state of a cell. scEGMM is a coherent method of cell type classification that includes no parameters and no clustering; however, it is comparable to existing machine learning-based methods in benchmarking studies and facilitates biological interpretation.

Download Full-text

Cell Type Classification and Unsupervised Morphological Phenotyping From Low-Resolution Images Using Deep Learning

Scientific Reports ◽

10.1038/s41598-019-50010-9 ◽

2019 ◽

Vol 9 (1) ◽

Cited By ~ 3

Author(s):

Kai Yao ◽

Nash D. Rochman ◽

Sean X. Sun

Keyword(s):

Cell Density ◽

Semantic Segmentation ◽

Cell Types ◽

Support Network ◽

Label Free ◽

Cell Type ◽

Morphological Heterogeneity ◽

A Cell ◽

Clustering Approach ◽

Type Classification

Abstract Convolutional neural networks (ConvNets) have proven to be successful in both the classification and semantic segmentation of cell images. Here we establish a method for cell type classification utilizing images taken with a benchtop microscope directly from cell culture flasks, eliminating the need for a dedicated imaging platform. Significant flask-to-flask morphological heterogeneity was discovered and overcome to support network generalization to novel data. Cell density was found to be a prominent source of heterogeneity even when cells are not in contact. For the same cell types, expert classification was poor for single-cell images and better for multi-cell images, suggesting experts rely on the identification of characteristic phenotypes within subsets of each population. We also introduce Self-Label Clustering (SLC), an unsupervised clustering method relying on feature extraction from the hidden layers of a ConvNet, capable of cellular morphological phenotyping. This clustering approach is able to identify distinct morphological phenotypes within a cell type, some of which are observed to be cell density dependent. Finally, our cell classification algorithm was able to accurately identify cells in mixed populations, showing that ConvNet cell type classification can be a label-free alternative to traditional cell sorting and identification.

Download Full-text

Single-Cell Entropy to Quantify the Cellular Order Parameter from Single-Cell RNA-Seq Data

Biophysical Reviews and Letters ◽

10.1142/s1793048020500010 ◽

2020 ◽

Vol 15 (01) ◽

pp. 35-49

Author(s):

Jingxin Liu ◽

You Song ◽

Jinzhi Lei

Keyword(s):

Single Cell ◽

Order Parameter ◽

Developmental Process ◽

Cell Types ◽

Gaussian Mixture ◽

Cell Type ◽

A Cell ◽

Molecular Components ◽

Different Cell Types ◽

Type Classification

The cell is the basic functional and biological unit of life, and a complex system that contains a huge number of molecular components. How can we quantify the macroscopic state of a cell from the microscopic information of these molecular components? This is a fundamental question to increase the understanding of the human body. The recent maturation of single-cell RNA sequencing (scRNA-seq) technologies has allowed researchers to gain information on the transcriptomes of individual cells. Although considerable progress has been made in terms of cell-type clustering over the past few years, there is no strong consensus about how to define a cell state from scRNA-seq data. Here, we present single-cell entropy (scEntropy) as an order parameter for cellular transcriptome profiles from scRNA-seq data. scEntropy is a straightforward parameter with which to define the intrinsic transcriptional state of a cell that can provide a quantity to measure the developmental process and to distinguish different cell types. The proposed scEntropy followed by Gaussian mixture model (scEGMM) provides a coherent method of cell-type classification that is simple, includes no parameters or clustering and is comparable to existing machine learning-based methods in benchmarking studies. The results of cell-type classification based on scEGMM are robust and easy to biologically interpret.

Download Full-text

Dendritic cells of the human epidermis: immuno-electron microscopic studies of la antigen reactivity

Proceedings, annual meeting, Electron Microscopy Society of America ◽

10.1017/s0424820100084090 ◽

1978 ◽

Vol 36 (2) ◽

pp. 644-645

Author(s):

G. Rowden ◽

M. G. Lewis ◽

T. M. Phillips

Keyword(s):

Dendritic Cells ◽

Langerhans Cells ◽

Cell Types ◽

Cell Type ◽

Electron Microscopic ◽

La Antigen ◽

Microscopic Studies ◽

A Cell ◽

C3 Receptors ◽

Antigen Reactivity

Langerhans cells of mammalian stratified squamous epithelial have proven to be an enigma since their discovery in 1868. These dendritic suprabasal cells have been considered as related to melanocytes either as effete cells, or as post divisional products. Although grafting experiments seemed to demonstrate the independence of the cell types, much confusion still exists. The presence in the epidermis of a cell type with morphological features seemingly shared by melanocytes and Langerhans cells has been especially troublesome. This so called "indeterminate", or " -dendritic cell" lacks both Langerhans cells granules and melanosomes, yet it is clearly not a keratinocyte. Suggestions have been made that it is related to either Langerhans cells or melanocyte. Recent studies have unequivocally demonstrated that Langerhans cells are independent cells with immune function. They display Fc and C3 receptors on their surface as well as la (immune region associated) antigens.

Download Full-text