Identification and Clustering of Event Patterns From In Vivo Multiphoton Optical Recordings of Neuronal Ensembles

In vivo multiphoton fluorescence microscopy allows imaging of cellular structures in brain tissue to depths of hundreds of micrometers and, when combined with the use of activity-dependent indicator dyes, opens the possibility of observing intact, functioning neural circuitry. We have developed tools for analyzing in vivo multiphoton data sets to identify responding structures and events in single cells as well as patterns of activity within the neural ensemble. Data were analyzed from populations of cerebellar Purkinje cell dendrites, which generate calcium-based complex action potentials. For image segmentation, active dendrites were identified using a correlation-based method to group covarying pixels. Firing events were extracted from dendritic fluorescence signals with a 95% detection rate and an 8% false-positive rate. Because an event that begins in one movie frame is sometimes not detected until the next frame, detection delays were compensated using a likelihood-based correction procedure. To identify groups of dendrites that tended to fire synchronously, a k-means-based procedure was developed to analyze pairwise correlations across the population. Because repeated runs of k-means often generated dissimilar clusterings, the runs were combined to determine a consensus cluster number and composition. This procedure, termed meta- k-means, gave clusterings as good as individual runs of k-means, was independent of random initial seeding, and allowed the exclusion of outliers. Our methods should be generally useful for analyzing multicellular activity recordings in a variety of brain structures.

Download Full-text

Target-Decoy MineR for determining the biological relevance of variables in noisy data sets

10.1101/2020.11.09.374181 ◽

2020 ◽

Author(s):

Cesaré Ovando-Vázquez ◽

Daniel Cázarez-García ◽

Robert Winkler

Keyword(s):

False Positive Rate ◽

Synthetic Data ◽

Noisy Data ◽

Machine Learning Algorithms ◽

Artificial Noise ◽

Data Sets ◽

Biological Relevance ◽

Different Types ◽

Positive Rate ◽

Sensitivity Specificity

AbstractMachine learning algorithms excavate important variables from biological big data. However, deciding on the biological relevance of identified variables is challenging. The addition of artificial noise, ‘decoy’ variables, to raw data, ‘target’ variables, enables calculating a false-positive rate (FPR) and a biological relevance probability (BRp) for each variable rank. These scores allow the setting of a cut-off for informative variables can be defined, depending on the required sensitivity/ specificity of a scientific question. We demonstrate the function of the Target-Decoy MineR (TDM) with synthetic data and with experimental metabolomics results. The Target-Decoy MineR is suitable for different types of quantitative data in tabular format. An implementation of the algorithm in R is freely available from https://bitbucket.org/cesaremov/targetdecoy_mining/.

Download Full-text

Excess False Positive Rates in Methods for Differential Gene Expression Analysis using RNA-Seq Data

10.1101/020784 ◽

2015 ◽

Cited By ~ 7

Author(s):

David M Rocke ◽

Luyao Ruan ◽

Yilun Zhang ◽

J. Jared Gossett ◽

Blythe Durbin-Johnson ◽

...

Keyword(s):

Linear Model ◽

False Positive ◽

Negative Binomial ◽

False Positive Rate ◽

Real Data ◽

False Positives ◽

P Value ◽

Data Sets ◽

Rna Seq ◽

Positive Rate

Motivation: An important property of a valid method for testing for differential expression is that the false positive rate should at least roughly correspond to the p-value cutoff, so that if 10,000 genes are tested at a p-value cutoff of 10−4, and if all the null hypotheses are true, then there should be only about 1 gene declared to be significantly differentially expressed. We tested this by resampling from existing RNA-Seq data sets and also by matched negative binomial simulations. Results: Methods we examined, which rely strongly on a negative binomial model, such as edgeR, DESeq, and DESeq2, show large numbers of false positives in both the resampled real-data case and in the simulated negative binomial case. This also occurs with a negative binomial generalized linear model function in R. Methods that use only the variance function, such as limma-voom, do not show excessive false positives, as is also the case with a variance stabilizing transformation followed by linear model analysis with limma. The excess false positives are likely caused by apparently small biases in estimation of negative binomial dispersion and, perhaps surprisingly, occur mostly when the mean and/or the dis-persion is high, rather than for low-count genes.

Download Full-text

A Learning-based Neural Network Model for the Detection and Classification of SQL Injection Attacks

Deep Learning and Neural Networks ◽

10.4018/978-1-7998-0414-7.ch026 ◽

2020 ◽

pp. 450-475

Author(s):

Naghmeh Moradpoor Sheykhkanloo

Keyword(s):

Neural Network ◽

Web Application ◽

Query Language ◽

False Positive Rate ◽

True Positive Rate ◽

Injection Technique ◽

Data Sets ◽

Injection Attacks ◽

Proposed Model ◽

Positive Rate

Structured Query Language injection (SQLi) attack is a code injection technique where hackers inject SQL commands into a database via a vulnerable web application. Injected SQL commands can modify the back-end SQL database and thus compromise the security of a web application. In the previous publications, the author has proposed a Neural Network (NN)-based model for detections and classifications of the SQLi attacks. The proposed model was built from three elements: 1) a Uniform Resource Locator (URL) generator, 2) a URL classifier, and 3) a NN model. The proposed model was successful to: 1) detect each generated URL as either a benign URL or a malicious, and 2) identify the type of SQLi attack for each malicious URL. The published results proved the effectiveness of the proposal. In this paper, the author re-evaluates the performance of the proposal through two scenarios using controversial data sets. The results of the experiments are presented in order to demonstrate the effectiveness of the proposed model in terms of accuracy, true-positive rate as well as false-positive rate.

Download Full-text

An Alternative Data Transformation Approach for ADA Cut Point Determination: Why Not Use a Weibull Transformation?

The AAPS Journal ◽

10.1208/s12248-021-00625-6 ◽

2021 ◽

Vol 23 (5) ◽

Author(s):

Gregor Jordan ◽

Roland F. Staack

Keyword(s):

Normal Distribution ◽

False Positive Rate ◽

Confirmatory Test ◽

Data Sets ◽

Screening Assay ◽

Healthy Human ◽

Data Set ◽

Cut Point ◽

Drug Candidates ◽

Positive Rate

AbstractThe testing of protein drug candidates for inducing the generation of anti-drug antibodies (ADA) plays a fundamental role in drug development. The basis of the testing strategy includes a screening assay followed by a confirmatory test. Screening assay cut points (CP) are calculated mainly based on two approaches, either non-parametric, when the data set does not appear normally distributed, or parametric, in the case of a normal distribution. A normal distribution of data is preferred and may be achieved after outlier exclusion and, if necessary, transformation of the data. The authors present a Weibull transformation and a comparison with a decision tree-based approach that was tested on 10 data sets (healthy human volunteer matrix, different projects). Emphasis is placed on a transformation calculation that can be easily reproduced to make it accessible to non-mathematicians. The cut point value and the effect on the false positive rate as well as the number of excluded samples of both methods are compared.

Download Full-text

Correlation of the Virological Response to Short-Term Maraviroc Monotherapy with Standard and Deep-Sequencing-Based Genotypic Tropism Prediction Methods

Antimicrobial Agents and Chemotherapy ◽

10.1128/aac.05857-11 ◽

2011 ◽

Vol 56 (3) ◽

pp. 1202-1207 ◽

Cited By ~ 15

Author(s):

A. Gonzalez-Serna ◽

R. A. McGovern ◽

P. R. Harrigan ◽

F. Vidal ◽

A. F. Y. Poon ◽

...

Keyword(s):

Deep Sequencing ◽

Virological Response ◽

Drug Exposure ◽

False Positive Rate ◽

Clinical Test ◽

Testing Methods ◽

Short Term ◽

Short Term Exposure ◽

Positive Rate

ABSTRACTGenotypic tropism testing methods are emerging as the first step before prescription of the CCR5 antagonist maraviroc (MVC) to HIV-infected patients in Europe. Studies validating genotypic tests have included other active drugs that could have potentially convoluted the effects of MVC. The maraviroc clinical test (MCT) is anin vivodrug sensitivity test based on the virological response to a short-term exposure to MVC monotherapy. Thus, our aim was to compare the results of genotypic tropism testing methods with the short-term virological response to MVC monotherapy. A virological response in the MCT was defined as a ≥1-log10decrease in HIV RNA or undetectability after 8 days of drug exposure. Seventy-three patients undergoing the MCT were included in this study. We used both standard genotypic methods (n= 73) and deep sequencing (n= 27) on MCT samples at baseline. For the standard methods, the most widely used genotypic algorithms for analyzing the V3 loop sequence, geno2pheno and PSSM, were used. For deep sequencing, the geno2pheno algorithm was used with a false-positive rate cutoff of 3.5. The discordance rates between the standard genotypic methods and the virological response were approximately 20% (including mostly patients without a virological response). Interestingly, these discordance rates were similar to that obtained from deep sequencing (18.5%). The discordance rates between the genotypic methods (tropism assays predictive of the use of the CCR5 coreceptor) and the MCT (in vivoMVC sensitivity assay) indicate that the algorithms used by genotypic methods are still not sufficiently optimized.

Download Full-text

DTI-CDF: a cascade deep forest model towards the prediction of drug-target interactions based on hybrid features

Briefings in Bioinformatics ◽

10.1093/bib/bbz152 ◽

2019 ◽

Cited By ~ 22

Author(s):

Yanyi Chu ◽

Aman Chandra Kaushik ◽

Xiangeng Wang ◽

Wei Wang ◽

Yufang Zhang ◽

...

Keyword(s):

Drug Target ◽

False Positive Rate ◽

Prediction Method ◽

Computational Prediction ◽

Data Sets ◽

Hybrid Features ◽

Positive Rate ◽

Deep Forest ◽

Wet Lab ◽

Test Sets

Abstract Drug–target interactions (DTIs) play a crucial role in target-based drug discovery and development. Computational prediction of DTIs can effectively complement experimental wet-lab techniques for the identification of DTIs, which are typically time- and resource-consuming. However, the performances of the current DTI prediction approaches suffer from a problem of low precision and high false-positive rate. In this study, we aim to develop a novel DTI prediction method for improving the prediction performance based on a cascade deep forest (CDF) model, named DTI-CDF, with multiple similarity-based features between drugs and the similarity-based features between target proteins extracted from the heterogeneous graph, which contains known DTIs. In the experiments, we built five replicates of 10-fold cross-validation under three different experimental settings of data sets, namely, corresponding DTI values of certain drugs (SD), targets (ST), or drug-target pairs (SP) in the training sets are missed but existed in the test sets. The experimental results demonstrate that our proposed approach DTI-CDF achieves a significantly higher performance than that of the traditional ensemble learning-based methods such as random forest and XGBoost, deep neural network, and the state-of-the-art methods such as DDR. Furthermore, there are 1352 newly predicted DTIs which are proved to be correct by KEGG and DrugBank databases. The data sets and source code are freely available at https://github.com//a96123155/DTI-CDF.

Download Full-text

DeepStreaks: identifying fast-moving objects in the Zwicky Transient Facility data with deep learning

Monthly Notices of the Royal Astronomical Society ◽

10.1093/mnras/stz1096 ◽

2019 ◽

Vol 486 (3) ◽

pp. 4158-4165 ◽

Cited By ~ 7

Author(s):

Dmitry A Duev ◽

Ashish Mahabal ◽

Quanzhi Ye ◽

Kushal Tirumala ◽

Justin Belicki ◽

...

Keyword(s):

Deep Learning ◽

Fast Moving ◽

Moving Objects ◽

False Positive Rate ◽

True Positive Rate ◽

Learning System ◽

Data Sets ◽

Wide Field ◽

Positive Rate ◽

Near Earth Objects

ABSTRACT We present DeepStreaks, a convolutional-neural-network, deep-learning system designed to efficiently identify streaking fast-moving near-Earth objects that are detected in the data of the Zwicky Transient Facility (ZTF), a wide-field, time-domain survey using a dedicated 47 deg2 camera attached to the Samuel Oschin 48-inch Telescope at the Palomar Observatory in California, United States. The system demonstrates a 96–98 per cent true positive rate, depending on the night, while keeping the false positive rate below 1 per cent. The sensitivity of DeepStreaks is quantified by the performance on the test data sets as well as using known near-Earth objects observed by ZTF. The system is deployed and adapted for usage within the ZTF Solar system framework and has significantly reduced human involvement in the streak identification process, from several hours to typically under 10 min per day.

Download Full-text

Comparative assessments of objective peak-detection algorithms. II. Studies in men

AJP Endocrinology and Metabolism ◽

10.1152/ajpendo.1988.254.1.e113 ◽

1988 ◽

Vol 254 (1) ◽

pp. E113-E119 ◽

Cited By ~ 3

Author(s):

R. J. Urban ◽

D. L. Kaiser ◽

E. van Cauter ◽

M. L. Johnson ◽

J. D. Veldhuis

Keyword(s):

False Positive ◽

False Positive Rate ◽

Peak Frequency ◽

Pulse Frequency ◽

Data Sets ◽

Empirical Distributions ◽

Detection Algorithms ◽

Sampling Intensity ◽

Positive Rate ◽

Simulated Series

The performances of eight currently available computerized pulse-detection algorithms were compared on signal-free noise and physiological luteinizing hormone (LH) time series. Signal-free noise was made to vary from 4 to 36% for Gaussian and empirical distributions. Physiological LH data were obtained by immunoassay of blood samples withdrawn every 5 min for 24 h in 8 healthy men, so that the data sets could be emended to simulate varying sampling intensities. Whenever possible, programs were tested at a presumptive 1% false-positive rate. In relation to signal-free noise, the Santen and Bardin program and its modification manifested elevated false-positive rates when the intraseries coefficients of variation increased. The Regional Dual-Threshold program yielded a 1% false-positive rate except on simulated series with high variance. The Cluster and Detect programs both approximated a 1% false-positive rate and the Ultra program approximated a 2.3% false-positive rate throughout the entire range of variance tested. In regard to physiological LH data, all algorithms disclosed a significant impact of sampling intensity on estimates of LH pulse frequency. Sampling-intensity dependent estimates of LH peak frequency by three of the eight programs (Ultra, Cluster, and Detect) were statistically indistinguishable from each other but distinct from the five other programs tested. Furthermore, when judged in relation to their ability to identify individual peaks, the three congruent programs were minimally distinguishable (McNemar's test). Rather, these programs identified the same particular peaks (as defined by concordance of peak maxima) at least 72% of the time.

Download Full-text

Relationship between simultaneously recorded spiking activity and fluorescence signal in GCaMP6 transgenic mice

10.1101/788802 ◽

2019 ◽

Cited By ~ 6

Author(s):

Lawrence Huang ◽

Ulf Knoblich ◽

Peter Ledochowitsch ◽

Jérôme Lecoq ◽

R. Clay Reid ◽

...

Keyword(s):

Calcium Imaging ◽

Pyramidal Neurons ◽

False Positive Rate ◽

Ground Truth ◽

Fluorescence Signal ◽

Spatiotemporal Resolution ◽

Detection Rates ◽

Single Spike ◽

Positive Rate

AbstractTwo-photon calcium imaging is often used with genetically encoded calcium indicators (GECIs) to investigate neural dynamics, but the relationship between fluorescence and action potentials (spikes) remains unclear. Pioneering work linked electrophysiology and calcium imaging in vivo with viral GECI expression, albeit in a small number of cells. Here we characterized the spikefluorescence transfer function in vivo of 91 layer 2/3 pyramidal neurons in primary visual cortex in four transgenic mouse lines expressing GCaMP6s or GCaMP6f. We found that GCaMP6s cells have spike-triggered fluorescence responses of larger amplitude, lower variability and greater single-spike detectability than GCaMP6f. Mean single-spike detection rates at high spatiotemporal resolution measured in our data was >70% for GCaMP6s and ~40-50% for GCaMP6f (at 5% false positive rate). These rates are estimated to decrease to 25-35% for GCaMP6f under generally used population imaging conditions. Our ground-truth dataset thus supports more refined inference of neuronal activity from calcium imaging.

Download Full-text

DTI-CDF: a CDF model towards the prediction of DTIs based on hybrid features

10.1101/657973 ◽

2019 ◽

Author(s):

Yan-Yi Chu ◽

Yu-Fang Zhang ◽

Wei Wang ◽

Xian-Geng Wang ◽

Xiao-Qi Shan ◽

...

Keyword(s):

Drug Target ◽

State Of The Art ◽

False Positive Rate ◽

Prediction Method ◽

Computational Prediction ◽

Data Sets ◽

Training Set ◽

Hybrid Features ◽

Positive Rate ◽

Deep Forest

AbstractDrug-target interactions play a crucial role in target-based drug discovery and exploitation. Computational prediction of DTIs has become a popular alternative strategy to the experimental methods for identification of DTIs of which are both time and resource consuming. However, the performances of the current DTIs prediction approaches suffer from a problem of low precision and high false positive rate. In this study, we aimed to develop a novel DTIs prediction method, named DTI-CDF, for improving the prediction precision based on a cascade deep forest model which integrates hybrid features, including multiple similarity-based features extracted from the heterogeneous graph, fingerprints of drugs, and evolution information of target protein sequences. In the experiments, we built five replicates of 10 fold cross-validations under three different experimental settings of data sets, namely, corresponding DTIs values of certain drugs (SD), targets (ST), or drug-target pairs (SP) in the training set are missed, but existed in the test set. The experimental results show that our proposed approach DTI-CDF achieved significantly higher performance than the state-of-the-art methods.

Download Full-text