scholarly journals Investigating Carboxysome Morphology Dynamics with a Rotationally Invariant Variational Autoencoder

2021 ◽  
Author(s):  
Miguel Fuentes-Cabrera ◽  
Jonathan K Sakkos ◽  
Daniel C. Ducat ◽  
Maxim Ziatdinov
Keyword(s):  
Protein Levels ◽  
Bacterial Microcompartments ◽  
Regulatory Processes ◽  
Learning Techniques ◽  
Endogenous Protein ◽  
Variational Autoencoder ◽  
Rotationally Invariant ◽  
Pcc 7942 ◽  
Microscopy Images ◽  
Over Time

Carboxysomes are a class of bacterial microcompartments that form proteinaceous organelles within the cytoplasm of cyanobacteria and play a central role in photosynthetic metabolism by defining a cellular microenvironment permissive to $CO_2$ fixation. Critical aspects of the assembly of the carboxysomes remain relatively unknown, especially with regard to the dynamics of this microcompartment. We have recently expressed an exogenous protease as a way of gaining control over endogenous protein levels, including carboxysomal components, in the model cyanobacterium \textit{Synechococcous elongatus} PCC 7942. By utilizing this system, proteins that compose the carboxysome can be tuned in real-time as a method to examine carboxysome dynamics. Yet, analysis of subtle changes in carboxysome morphology with microscopy remains a low-throughput and subjective process. Here we use deep learning techniques, specifically a Rotationally Invariant Variational Autoencoder (rVAE), to analyze the fluorescence microscopy images and quantitatively evaluate how carboxysome shell remodelling impacts trends in the morphology of the microcompartment over time. We find that rVAEs are able to assist in the quantitative evaluation of changes in carboxysome location, shape, and size over time. We propose that rVAEs may be a useful tool to accelerate the analysis of carboxysome assembly and dynamics in response to genetic or environmental perturbation, and may be more generally useful to probe regulatory processes involving a broader array of bacterial microcompartments.

scholarly journals Training a deep learning model for single-cell segmentation without manual annotation

2021 ◽  
Vol 11 (1) ◽  
Author(s):  
Nizam Ud Din ◽  
Ji Yu
Keyword(s):  
Machine Learning ◽  
Training Data ◽  
Machine Learning Techniques ◽  
Cell Segmentation ◽  
Learning Techniques ◽  
Signal Characteristics ◽  
Human Operators ◽  
Deep Learning Model ◽  
Microscopy Images ◽  
Segmentation Models

AbstractAdvances in the artificial neural network have made machine learning techniques increasingly more important in image analysis tasks. Recently, convolutional neural networks (CNN) have been applied to the problem of cell segmentation from microscopy images. However, previous methods used a supervised training paradigm in order to create an accurate segmentation model. This strategy requires a large amount of manually labeled cellular images, in which accurate segmentations at pixel level were produced by human operators. Generating training data is expensive and a major hindrance in the wider adoption of machine learning based methods for cell segmentation. Here we present an alternative strategy that trains CNNs without any human-labeled data. We show that our method is able to produce accurate segmentation models, and is applicable to both fluorescence and bright-field images, and requires little to no prior knowledge of the signal characteristics.

Potassium channel KCNJ15 is required for histamine-stimulated gastric acid secretion

2015 ◽  
Vol 309 (4) ◽  
pp. C264-C270 ◽  
Author(s):  
Jianye Yuan ◽  
Wensheng Liu ◽  
Serhan Karvar ◽  
Susan S. Baker ◽  
Wenjun He ◽  
...  
Keyword(s):  
Acid Secretion ◽  
Gastric Acid Secretion ◽  
Gastric Acid ◽  
Apical Membrane ◽  
Imaging System ◽  
Parietal Cells ◽  
Live Cells ◽  
Protein Levels ◽  
Endogenous Protein ◽  
Luminal Side

Gastric acid secretion is mediated by the K+-dependent proton pump (H+,K+-ATPase), which requires a continuous supply of K+ at the luminal side of the apical membrane. Several K+ channels are implicated in gastric acid secretion. However, the identity of the K+ channel(s) responsible for apical K+ supply is still elusive. Our previous studies have shown the translocation of KCNJ15 from cytoplasmic vesicles to the apical membrane on stimulation, indicating its involvement in gastric acid secretion. In this study, the stimulation associated trafficking of KCNJ15 was observed in a more native context with a live cell imaging system. KCNJ15 molecules in resting live cells were scattered in cytoplasm but exhibited apical localization after stimulation. Furthermore, knocking down KCNJ15 expression with a short hairpin RNA adenoviral construct abolished histamine-stimulated acid secretion in rabbit primary parietal cells. Moreover, KCNJ15, like H+,K+-ATPase, was detected in all of the parietal cells by immunofluorescence staining, whereas only about half of the parietal cells were positive for KCNQ1 under the same condition. Consistently, the endogenous protein levels of KCNJ15, analyzed by Western blotting, were higher than those of KCNQ1 in the gastric mucosa of human, mouse, and rabbit. These results provide evidence for a major role of KCNJ15 in apical K+ supply during stimulated acid secretion.

scholarly journals A Semi-Supervised Learning Approach for Tackling Twitter Spam Drift

2019 ◽  
Vol 18 (02) ◽  
pp. 1950010 ◽  
Author(s):  
Niddal Imam ◽  
Biju Issac ◽  
Seibu Mary Jacob
Keyword(s):  
Machine Learning ◽  
Supervised Learning ◽  
Research Community ◽  
Machine Learning Techniques ◽  
Spam Detection ◽  
Learning Approach ◽  
New Approach ◽  
Detection Systems ◽  
Learning Techniques ◽  
Over Time

Twitter has changed the way people get information by allowing them to express their opinion and comments on the daily tweets. Unfortunately, due to the high popularity of Twitter, it has become very attractive to spammers. Unlike other types of spam, Twitter spam has become a serious issue in the last few years. The large number of users and the high amount of information being shared on Twitter play an important role in accelerating the spread of spam. In order to protect the users, Twitter and the research community have been developing different spam detection systems by applying different machine-learning techniques. However, a recent study showed that the current machine learning-based detection systems are not able to detect spam accurately because spam tweet characteristics vary over time. This issue is called “Twitter Spam Drift”. In this paper, a semi-supervised learning approach (SSLA) has been proposed to tackle this. The new approach uses the unlabeled data to learn the structure of the domain. Different experiments were performed on English and Arabic datasets to test and evaluate the proposed approach and the results show that the proposed SSLA can reduce the effect of Twitter spam drift and outperform the existing techniques.

scholarly journals Popular music lyrics and musicians’ gender over time: A computational approach

2019 ◽  
pp. 030573561987160 ◽  
Author(s):  
Manuel Anglada-Tort ◽  
Amanda E Krause ◽  
Adrian C North
Keyword(s):  
Machine Learning ◽  
Popular Music ◽  
Machine Learning Techniques ◽  
Mixed Effect ◽  
Gender Distribution ◽  
Learning Techniques ◽  
Inflection Points ◽  
The Uk ◽  
Music Lyrics ◽  
Over Time

The present study investigated how the gender distribution of the United Kingdom’s most popular artists has changed over time and the extent to which these changes might relate to popular music lyrics. Using data mining and machine learning techniques, we analyzed all songs that reached the UK weekly top 5 sales charts from 1960 to 2015 (4,222 songs). DICTION software facilitated a computerized analysis of the lyrics, measuring a total of 36 lyrical variables per song. Results showed a significant inequality in gender representation on the charts. However, the presence of female musicians increased significantly over the time span. The most critical inflection points leading to changes in the prevalence of female musicians were in 1968, 1976, and 1984. Linear mixed-effect models showed that the total number of words and the use of self-reference in popular music lyrics changed significantly as a function of musicians’ gender distribution over time, and particularly around the three critical inflection points identified. Irrespective of gender, there was a significant trend toward increasing repetition in the lyrics over time. Results are discussed in terms of the potential advantages of using machine learning techniques to study naturalistic singles sales charts data.

Comparison of Serum Free Light Chain Levels and 24 Hour Urinary Monoclonal Protein Secretion in Patients with Myeloma: Concordant Changes with Response to Therapy.

Blood ◽  
2006 ◽  
Vol 108 (11) ◽  
pp. 5064-5064 ◽  
Author(s):  
Shaji Kumar ◽  
S. Vincent Rajkumar ◽  
Matthew Plevak ◽  
Robert A. Kyle ◽  
Jerry A. Katzmann ◽  
...  
Keyword(s):  
Light Chain ◽  
Protein Level ◽  
Free Light Chain ◽  
Response To Therapy ◽  
M Protein ◽  
Serum Free ◽  
Protein Levels ◽  
Serum Free Light Chain ◽  
Data Points ◽  
Over Time

Abstract Background: The measurement of monoclonal (M) protein in the serum and urine is critical for response assessment and disease evaluation in patients with multiple myeloma (MM). The serum free light chain (FLC) assay offers a new and sensitive method of assessing response to therapy. An important question that has not been adequately addressed is the correlation between 24 hour urine M protein levels and serum FLC measurements, and the extent to which response to therapy estimated using the FLC assay correlates with that assessed using the 24 hour urine M protein level. Methods: A total of 2194 sets of data, with simultaneous UPEP and serum FLC measurement, were studied. These included 752 unique patients, with individual patients having 1–23 paired assessments over time. FLC estimation was carried out using the serum FLC assay (Freelite; The Binding Site Limited, UK) performed on a Dade-Behring Nephelometer. Based on the established reference range, kappa/lambda FLC ratio <0.26 or >1.65 were defined as abnormal indicating the presence of monoclonal lambda and kappa FLC, respectively. The monoclonal light chain isotype was considered the involved FLC isotype, and the opposite light chain type as the uninvolved FLC type. The Urine M protein by UPEP was compared to the serum levels of the involved light chain using Spearman Rank Correlation. For comparisons in individual patients over time, those with at least 10 measurements each were studied. Results: The median involved FLC level in patients with an undetectable urine M protein was 2.3 mg/dl compared to 32.2 mg/dL among those with a detectable urine M protein (P<0.001). Among the 1676 points with an abnormal FLC ratio, only 75% had an M protein detected in the urine, P < 0.001. Conversely, among patients with a positive urine M-protein, 91% had an abnormal FLC ratio. When all the 2194 data points were considered together, there was a significant correlation between the urine M protein level and the FLC levels (FLC level calculated as the difference between involved and uninvolved levels), rho=0.763, P < 0.001. The correlation did not change when patients with a serum creatinine of over 2.5 were excluded. The correlation between FLC levels and urinary M protein can be affected by several factors such as renal function that will differ across patients. Therefore, we examined whether the correlation between the two variables is stronger when the variations introduced by inter-patient differences in the relationship between the two variables are eliminated. In order to do this, we studied individual patients on whom multiple data points over time were available. One patient who had the maximum number of paired assessments (23 pairs) of serum FLC level and urinary M protein; the correlation between the two variables over time was highly significant, rho 0.981, p<0.001. Similarly 26 other patients who had measurable urine M protein levels in whom 10 nor more paired observations over time were available, also showed significant correlations, rho, range 0.726–0.981, p<0.01. Conclusion: There is a significant correlation between urine M-protein and serum free light chain across patients and the correlation is stronger in individual patients in whom the effect of inter-patient variation in other confounding factors can be eliminated. These data if confirmed in a clinical trial setting would support the use of serum FLC levels instead of urinary M protein measurements to assess response to therapy.

scholarly journals Exploring peptide capture by anti-protein antibodies for LC-MS-based biomarker determination

2018 ◽  
Author(s):  
Maren Levernæs ◽  
Bassem Farhat ◽  
Inger Oulie ◽  
Sazan S. Abdullah ◽  
Elisabeth Paus ◽  
...  
Keyword(s):  
Protein Extraction ◽  
High Sensitivity ◽  
Intact Protein ◽  
Protein Biomarkers ◽  
Promising Technique ◽  
Intact Proteins ◽  
Protein Levels ◽  
Endogenous Protein ◽  
Sensitivity And Selectivity

<p>Immunocapture LC-MS/MS is a promising technique to ensure high sensitivity and selectivity of low-abundant protein biomarkers. For this purpose, the use of monoclonal antibodies (mAb) is especially attractive as they are renewable reagents that can be standardized. In this article we investigated the possibility of using mAbs developed against intact proteins (anti-protein antibodies) to capture proteotypic epitope peptides. Three mAbs were tested, and all selectively extracted proteotypic epitope peptides from a complex sample. Compared to intact protein extraction, this concept which we call peptide capture by anti-protein antibodies provided cleaner extracts, which further improved the sensitivity. Analysis of three patient samples demonstrated that p can be used for the determination of different endogenous protein levels. </p><p></p>

scholarly journals Unsupervised deep learning method for cell segmentation

2021 ◽  
Author(s):  
Nizam Ud Din ◽  
Ji Yu
Keyword(s):  
Machine Learning ◽  
Training Data ◽  
Machine Learning Techniques ◽  
Cell Segmentation ◽  
Learning Techniques ◽  
Signal Characteristics ◽  
Unsupervised Deep Learning ◽  
Human Operators ◽  
Microscopy Images ◽  
Segmentation Models

Advances in the artificial neural network have made machine learning techniques increasingly more important in image analysis tasks. More recently, convolutional neural networks (CNN) have been applied to the problem of cell segmentation from microscopy images. However, previous methods used a supervised training paradigm in order to create an accurate segmentation model. This strategy requires a large amount of manually labeled cellular images, in which accurate segmentations at pixel level were produced by human operators. Generating training data is expensive and a major hindrance in the wider adoption of machine learning based methods for cell segmentation. Here we present an alternative strategy that uses unsupervised learning to train CNNs without any human-labeled data. We show that our method is able to produce accurate segmentation models. More importantly, the algorithm is applicable to both fluorescence and bright-field images, requiring no prior knowledge of signal characteristics and requires no tuning of parameters.

scholarly journals Quality of bovine colostrum and its relation to genetics, management, physiology and its freezing

2020 ◽  
pp. e1465
Author(s):  
Ana Luiza Schogor ◽  
Patricia Glombowsky ◽  
Fabiana Both ◽  
Beatriz Danieli ◽  
Fernanda Rigon ◽  
...  
Keyword(s):  
Holstein Cows ◽  
Total Serum Protein ◽  
Total Serum ◽  
Freezing Process ◽  
Plasma Protein Concentration ◽  
Protein Levels ◽  
The Mean ◽  
Over Time ◽  
Process Material

Objective. The aims of this study were to assess whether colostrum quality is modified by genetic, physiological and management characteristics in the pre-partum period, as well as evaluate whether quality and composition of colostrum is altered in the freezing process. Material and methods. In the experiment I, colostrum and blood samples of 35 cows (18 Holstein and 17 Jerseys) were collected. In the experiment II, six colostrum samples of Holstein cows were collected and frozen during 60 days. Results. The mean immunoglobulin (Ig) concentration was 77.65 mg/ml to Jersey and 82.77 mg/ml to Holstein. The genetic, parturition order, and the interaction between these factors were no significant on IgG concentration in the colostrum. Also, it was observed an effect genetic of cow in the weight on calf at birth and on three days of age (p<0.0001). Regarding transmission of calf passive immunity, no effects of cow breed and calving order were observed on plasma protein concentration of calf, as well as after three days of freezing. Calves of Holstein (83%) and Jersey (82%) breed showed total serum protein levels above 5.5 g/dL. Holstein cows housed in individual paddocks with diet supplementation provided better quality of colostrum (93.57 mg Ig/mL). Over time, the percentage of fat reduced at freezing, that reduced over time (p<0.05) in Experiment II. Conclusions. The pre-partum management exerts influence on colostrum quality, and the freezing not interfere on centesimal and immunological quality of colostrum, with exception the fat, that decrease along the time.

scholarly journals SISUA: Semi-Supervised Generative Autoencoder for Single Cell Data

2019 ◽  
Author(s):  
Trung Ngo Trong ◽  
Roger Kramer ◽  
Juha Mehtonen ◽  
Gerardo González ◽  
Ville Hautamäki ◽  
...  
Keyword(s):  
Single Cell ◽  
Network Architecture ◽  
Surface Protein ◽  
Protein Quantification ◽  
Additional Information ◽  
Protein Levels ◽  
Variational Autoencoder ◽  
Cell Gene Expression ◽  
Cell Phenotypes ◽  
Cell Data

ABSTRACTSingle-cell transcriptomics offers a tool to study the diversity of cell phenotypes through snapshots of the abundance of mRNA in individual cells. Often there is additional information available besides the single cell gene expression counts, such as bulk transcriptome data from the same tissue, or quantification of surface protein levels from the same cells. In this study, we propose models based on the Bayesian generative approach, where protein quantification available as CITE-seq counts from the same cells are used to constrain the learning process, thus forming a semi-supervised model. The generative model is based on the deep variational autoencoder (VAE) neural network architecture.

Sign in / Sign up

Export Citation Format

Share Document