Violin Plot Research Articles

Visualization of Data Mining Techniques for the Prediction of Breast Cancer with High Accuracy Rates

Breast cancer is one of the leading causes of death in women worldwide. Around one in 30 women are affected by breast cancer. Mammography has helped in detecting breast cancer in the early stages which have reduced mortality. The diagnosis of breast cancer is dependent on a variety of parameters. In this paper, we aim to create the best model for predicting breast cancer through preprocessing, feature extraction, data visualization and prediction using breast cancer data. Various visualization techniques like violin plot, grid plot, swarm plot and heat plot were utilized for proper feature extraction which has improved the accuracy of our results. For the purpose of prediction, we have used algorithms like the random forest, decision tree with single and multiple predictors, along with the commonly used statistical model, logistic regression model. We have also relied on 5-fold cross-validation methods to measure the unbiasedness of the prediction models for performance reasons. An analysis of the models was carried out and the best model was selected based on its accuracy. The results showcased that the random forest model provided an accuracy rate of 94.724% with decent 5-fold cross-validation, followed by the decision tree model which had an accuracy rate of 100% with poor 5-fold cross-validation. This was followed by the logistic regression model which had an accuracy rate of 88.442% with a low 5-fold cross-validation score.

Focus on software, data acquisition, and instrumentation

Focus on software, data acquisition, and instrumentation

Complementary data-intelligence model for river flow simulation

Despite of diverse progressions in hydrological modeling techniques, the necessity of a robust, accurate, reliable, and trusted expert system for real-time stream flow prediction still exists. The intention of the present study is to establish a new complementary data-intelligence (DI) model called wavelet extreme learning machine (WA-ELM) for forecasting river flow in a semi-arid environment. The monthly river flow data for the period 1991-2010 is used to calibrate and validate the applied predictive model, developed using antecedent flow data as predictor. The prediction efficiency of the developed WA-ELM model is validated against stand-alone ELM model. The performance of the models is diagnosed using multiple statistical metrics and graphical analysis visualization. The results reveal that incorporation of data pre-processing wavelet approach with ELM model enhances the river flow predictability. In quantitative term, the root-mean-square error (RMSE) and mean absolute error (MAE) measurements are reduced by 65% and 67% using WA-ELM over ELM model, respectively. The Taylor diagram reveals much closer proximity and the Violin plot shows similar distribution of WA-ELM simulated river flow to the observed river flow compared to stand-alone ELM simulated river flow. The hybridization of wavelet decomposition method with ELM model improves the ability of ELM model to extract the required information for modeling the non-stationary and high stochastic river flow pattern. Overall, the study reveals that WA-ELM can be a reliable methodology for modeling river flow in semi-arid environment and for different regimes (i.e., low-, medium- and high-flow).

Feasibility of representing adherence to blood glucose monitoring through visualizations: A pilot survey study among healthcare workers

BackgroundMeasuring adherence to processes is one of the established ways to quantify the quality of healthcare. Providing timely feedback to healthcare workers on the level of adherence can improve process measures. However, it is challenging to present data on adherence to repetitive time-sensitive tasks in a clear manner. ObjectivesWe used inpatient glucose monitoring as a test case to explore the feasibility of using visualizations to communicate adherence to repetitive scheduled tasks to healthcare workers. MethodsWe selected four candidate plots that represented distribution across time: histogram, probability density function plot (pdf plot), violin plot and cumulative density function plot (cdf plot). Doctors and nurses involved in inpatient diabetes care in a tertiary hospital were invited to complete a self-administered questionnaire that measured self-reported baseline knowledge, performance, and perception towards the visualizations. Performance was assessed by determining if a participant was able to correctly identify visualizations representing protocol adherence. We also assessed the perception of usability of these visualizations for monitoring protocol adherence. Binomial regression models were used to identify factors associated with overall performance and perception. Logistic regression models with generalized estimating equation were used to compare performance and perception between visualizations, and identify effect modifiers. ResultsA total of 57 doctors and nurses completed the questionnaire. Participants were most familiar with histogram (87.7%), followed by cdf plot (61.4%), pdf plot (40.4%), and violin plot (7%). However, the percentages of participants who identified non-adherence using these plots were generally lower, ranging from 29.8% to 40.4%. Participants’ perception of usability ranged from 14% to 17.5% across these visualizations. More favorable perceptions were found among participants with baseline knowledge for two or more visualizations (adjusted odds ratio: 3.21; 95%CI: 1.29, 7.96; p-value: 0.012) and having identified two or more non-adherent visualizations (adjusted odds ratio: 4.23; 95%CI: 1.95, 9.16; p-value: < 0.001). ConclusionsAdherence to repetitive time-sensitive tasks can be presented in the form of visualizations. However, nurses’ and doctors’ knowledge and understanding of these visualizations are generally poor. This may influence their perception of usability of these plots. Therefore, these visualizations need to be implemented in tandem with training on their interpretation, to enhance the usefulness of these plots in motivating quality improvement.

SinaPlot: An Enhanced Chart for Simple and Truthful Representation of Single Observations Over Multiple Classes

ABSTRACTRecent developments in data-driven science have led researchers to integrate data from several sources, over diverse experimental procedures, or databases. This alone poses a major challenge in truthfully visualizing data, especially when the number of data points varies between classes. To aid the representation of datasets with differing sample size, we have developed a new type of plot overcoming limitations of current standard visualization charts. SinaPlot is inspired by the strip chart and the violin plot and operates by letting the normalized density of points restrict the jitter along the x-axis. The plot displays the same contour as a violin plot but resembles a simple strip chart for a small number of data points. By normalizing jitter over all classes, the plot provides a fair representation for comparison between classes with a varying number of samples. In this way, the plot conveys information of both the number of data points, the density distribution, outliers and data spread in a very simple, comprehensible, and condensed format. The package for producing the plots is available for R through the CRAN network using base graphics package and as geom for ggplot through ggforce. We also provide access to a web-server accepting excel sheets to produce the plots (http://servers.binf.ku.dk:8890/sinaplot/).

Which method is more suitable for advanced gastric cancer with enlarged lymph nodes, laparoscopic radical gastrectomy or open gastrectomy?

The oncologic outcome of laparoscopic radical gastrectomy (LG) with open radical gastrectomy (OG) for gastric cancer with preoperative enlarged lymph nodes (LNs) remains ambiguous. The long-term outcomes of LG (n=855) versus OG (n=154) in gastric adenocarcinoma patients were analyzed retrospectively. Patients were stratified according to enlarged (>10mm) and small (≤10mm) LNs (ELN and SLN) based on the long-axis diameter of the LNs. The violin plot indicates that the distribution of ELN size was similar between two groups. Survival curves demonstrated that the overall survival (OS) in LG is enhanced compared with OG (p=0.034). A stratified analysis revealed that the OS was better in the LG group compared with the OG group for patients with ELNs (p=0.038). In a forest map analysis, the actual 3-year OS rate for LG was significantly increased compared with OG in enlarged suprapancreatic LN (ESLNs) patients. Stratified analysis based on different diameters of ESLNsrevealed that the actual 3-year OS and cancer-specific survival (CSS)rate for LG was significant increased compared to OG with a diameter from 1.0 to 1.9cm for pII or pIII stage patients. However, the actual 3-year OS and CSS rate for LG was significantly reduced compared with OG when the size exceeded 2.5cm. For advanced gastric cancer with an ESLN diameter from 1.0 to 1.9cm, LG couldbe chosen preferentially; nevertheless, LG is not safe if ESLN are>2.5 cm.

Single-cell protein-mRNA correlation analysis enabled by multiplexed dual-analyte co-detection

We have investigated the correlation between proteins and mRNAs in single cells employing an integrated workflow for dual-analyte co-detection. This is achieved by combining the oligo extension reaction (OER), which converts protein levels to DNA levels, with reverse transcription for mRNA detection. Unsupervised gene expression profiling analysis, including principal component analysis and hierarchical clustering, revealed different aspects of the protein-mRNA relationship. Violin plot analysis showed that some genes exhibited similar distribution patterns for proteins and mRNAs. We also demonstrate that cells can be separated into subpopulations based on their protein-mRNA expression profiles, and that different subpopulations have distinct correlation coefficient values. Our results demonstrated that integrated investigations of mRNA and protein levels in single cells allows comprehensive analysis not attainable at bulk levels.

Resiliency in Education: The Case of Foreign Language Teachers

The article covers the basics of resilience and the factors promoting sustainability within the field of education. With escalating demands on teachers in terms of increasing social pressures and expectations regarding quality of pedagogical provision, ’resilience’ is a topical area of research. The main objectives of the study include ascertaining how prone to professional stress are educators compared to other professionals, pointing out the areas where the major mismatches lie, and revealing the factors influencing professional adjustment in the sphere of foreign language teaching. The correlation and interdependence of resilience, emotional intelligence, social support and other individual factors have been studied using a valid sample of specialists in language teaching: resilience was considered one of the central, individual protective factors, as well as a serious prerequisite for a successful and fulfilling life. The sample of the study comprised three groups: first, senior students (future teachers); second, university academics (senior professors) and thirdly, university junior specialists. To analyze the results, statistical and correlation methods were used: different types of interdependence were indicated through regression analysis, the Kruskal–Wallis test, violin plot, and others.

Edaphic niches of metallophytes from southeastern Democratic Republic of Congo: Implications for post-mining restoration

In southeastern D. R. Congo, about 550 metallophytes grow on soils with high copper and cobalt concentrations, 57 of which are endemics to these metalliferous environments. About 70% of those endemics are considered threatened by destruction of habitats through mining activities. To provide guidelines for future restoration programs, the edaphic ecological niches of eight endemic metallophytes (i.e. copper endemics) were studied using a pragmatic sampling method adapted for urgent conservation needs. Niches were modelled using violin plot along Cu, Co and C:N gradients representing the two main independent edaphic gradients among nine edaphic variables (C, N, C:N, K, P, pH, Co, Cu, and Mn). Copper endemics presented distinct edaphic niches along the copper and cobalt gradients, but differentiation was lower along the C:N gradient. In addition, edaphic elements presented covariations among them and metalliferous soils had higher nutrient and element content compared to the non-metalliferous soils of the region dominated by the Miombo woodland. The complexity of the soil composition and the edaphic niches of copper endemics revealed an important challenge in the implementation of the species conservation and the habitat restoration strategies of post-mining sites.

Visualization of Skewed Data

After discussing the main characteristics of the histogram and of a number of variations in the boxplot, this work presents a visualization tool specifically tailored to deal with skewed data. The idea is to use various types of boxplots (the classical one, which is tuned for skewness of the data, the shifting boxplot, and the box-percentile plot), the violin plot, and the histogram with a nonparametric estimate of the density overlay. The plots are presented in such a way that they facilitate the extraction of additional information from each one. We show that a good deal of information can be extracted from the inspection of the output using example data from synthetic aperture radar images. We provide an implementation in R based on functions already available.

Abstract 5588: Epithelial-to-mesenchymal markers of circulating tumor cells for detection of castration-resistant prostate cancer

Abstract Approximately 30% of prostate cancer (PCa) patients who receive androgen-deprivation therapy experience disease progression, including bone metastasis, in 18-36 months. Clinical diagnosis of this castration resistance is primarily based on a continuous rise in serum PSA levels during therapy. However, the underlying disease progression predates the clinical onset of castration resistance by many months. The clinical challenge is to distinguish indolent vs aggressive cancer for better disease management. During the prostate cancer progression, malignant circulating tumor cells (CTC) shed from the primary site into the bloodstream through epigenetic modifications and an epithelial-to-mesenchymal transition (EMT). Evidence suggests that increased CTCs are present in the blood of castration-resistant PCa patients and can be obtained through routine phlebotomy. EMT changes of CTCs present a great potential for disease prognosis, treatment stratification and subsequent disease monitoring. We have recently established the combined microfiltration-micromanipulation system (CM2S) to enrich prostate CTCs from blood as published in The Prostate. Our preliminary studies based on a relatively small sample size show that CTCs isolated from advanced PCa patients often lose the typical features of prostate epithelial cells and display incremental EMT-related gene expression signatures, higher elasticity and smoother membrane features. Furthermore, incremental expression of these genes and particular nanomechanical features in CTCs are associated with castration-resistant and metastatic PCa. In this study, we would like expand the sample size to clinically develop this methodology. CTCs were isolated from patients' blood (∼15 ml) using microfiltration system (ScreenCell). We analyzed castration-sensitive (CS) and castration-resistant (CR) PCa cells and CTCs using high throughput microfluidic single-cell RT-PCR. The data were subject to hierarchical clustering, violin plot and Ingenuity Pathway Analysis (IPA). Increased expression of EMT genes was found in CR PCa cells and CTCs as compared to CS counterparts. IPA indicated that these EMT genes are closely related to AKT, β-catenin, Myc and NFκB pathways. Some of CTCs isolated from patients were subject to nanomechanical and nanochemical analysis using PeakForce QNM Catalyst atomic force microscopy (AFM) (Bruker). CTCs of CR patients are about 3 fold more elastic and 7 fold more adherent than CTCs of CS patients. Additionally, CTCs of CR patients are about 3 fold more deformable than CTCs of CS patients. In conclusion, we confirmed our previous study with a larger patient sample size. The increased expression of EMT-related genes and nanomechanical and nanochemical phenotypes in CTCs are potential biomarkers for detection of the castration-resistant PCa, treatment stratification and subsequent disease monitoring. Citation Format: Chun-Liang Chen, Pawel Osmulski, Devalingam Mahalingam, Aaron M. Horning, Rohit R. Jadhav, Anna D. Louie, Chiou-Miin Wang, Tim H.-M. Huang. Epithelial-to-mesenchymal markers of circulating tumor cells for detection of castration-resistant prostate cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 5588. doi:10.1158/1538-7445.AM2014-5588

BoxPlotR: a web tool for generation of box plots

To the Editor In biomedical research, it is often necessary to compare multiple data sets with different distributions. The bar plot, or histogram, is typically used to compare data sets on the basis of simple statistical measures, usually the mean with s.d. or s.e.m. However, summary statistics alone may fail to convey underlying differences in the structure of the primary data (Fig. 1a), which may in turn lead to erroneous conclusions. The box plot, also known as the box-and-whisker plot, represents both the summary statistics and the distribution of the primary data. The box plot thus enables visualization of the minimum, lower quartile, median, upper quartile and maximum of any data set (Fig. 1b). The first documented description of a box plot–like graph by Spear1 defined a range bar to show the median and interquartile range (IQR, or middle 50%) of a data set, with whiskers extended to minimum and maximum values. The most common implementation of the box plot, as defined by Tukey2, has a box that represents the IQR, with whiskers that extend 1.5 times the IQR from the box edges; it also allows for identification of outliers in the data set. Whiskers can also be defined to span the 95% central range of the data3. Other variations, including bean plots4 and violin plots, reveal additional details of the data distribution. These latter variants are less statistically informative but allow better visualization of the data distribution, such as bimodality (Fig. 1b), that may be hidden in a standard box plot. Figure 1 Data visualization with box plots Despite the obvious advantages of the box plot for simultaneous representation of data set and statistical parameters, this method is not in common use, in part because few available software tools allow the facile generation of box plots. For example, the standard spreadsheet tool Excel is unable to generate box plots. Here we describe an open-source application, called BoxPlotR, and an associated web portal that allow rapid generation of customized box plots. A user-defined data matrix is uploaded as a file or pasted directly into the application to generate a basic box plot with options for additional features. Sample size may be represented by the width of each box in proportion to the square root of the number of observations5. Whiskers may be defined according to the criteria of Spear1, Tukey2 or Altman3. The underlying data distribution may be visualized as a violin or bean plot or, alternatively, the actual data may be displayed as overlapping or nonoverlapping points. The 95% confidence interval that two medians are different may be illustrated as notches defined as ±(1.58 × IQR/√n) (ref. 5). There is also an op on to plot the sample means and their confidence intervals. More complex statistical comparisons may be required to ascertain significance according to the specific experimental design6. The output plots may be labeled; customized by color, dimensions and orientation; and exported as publication-quality .eps, .pdf or .svg files. To help ensure that generated plots are accurately described in publications, the application generates a description of the plot for incorporation into a figure legend. The interactive web application is written in R (ref. 7) with the R packages shiny, beanplot4, vioplot, beeswarm and RColorBrewer, and it is hosted on a shiny server to allow for interactive data analysis. User data are held only temporarily and discarded as soon as the session terminates. BoxPlotR is available at http://boxplot.tyerslab.com/ and may be downloaded to run locally or as a virtual machine for VMware and VirtualBox.

Graphical and numerical descriptive analysis: Exploratory tools applied to Vietnamese data

This case study covers several exploratory data analysis ideas, the histogram and boxplot, kernel density estimates, the recently introduced bagplot – a two-dimensional extension of the boxplot – as well as the violin plot, which combines a boxplot with a density shape plot. We apply these ideas and demonstrate how to interpret the output from these tools in the context of data on living standards in Vietnam. The level of the presentation is suitable for an upper-level undergraduate or beginning graduate course in applied statistics. We use data from the Vietnam Living Standards Survey of 1998 (VLSS98) and from the 2000 Vietnam statistical yearbook, the statistical package Stata, and special programs provided by the authors who introduced the bagplot and the violin plot.

Graphical and Numerical Descriptive Analysis: Exploratory Tools Applied to Vietnamese Data

This case study covers several exploratory data analysis ideas, the histogram and boxplot, kernel density estimates, the recently introduced bagplot - a two-dimensional extension of the boxplot - as well as the violin plot, which combines a boxplot with a density shape plot. We apply these ideas and demonstrate how to interpret the output from these tools in the context of data on living standards in Vietnam. The level of the presentation is suitable for an upper-level undergraduate or beginning graduate course in applied statistics. We use data from the Vietnam Living Standards Survey of 1998 (VLSS98) and from the 2000 Vietnam statistical yearbook, the statistical package Stata, and special programs provided by the authors who introduced the bagplot and the violin plot.Supplemental data for this article can be accessed on the publisher's website.

Violin Plots: A Box Plot-Density Trace Synergism

Many modifications build on Tukey's original box plot. A proposed further adaptation, the violin plot, pools the best statistical features of alternative graphical representations of batches of data. It adds the information available from local density estimates to the basic summary statistics inherent in box plots. This marriage of summary statistics and density shape into a single plot provides a useful tool for data analysis and exploration.