ABSTRACT
MOTIVATION: Meta-analysis methods have been widely used to combine results from multiple clinical or genomic studies to increase statistical powers and ensure robust and accurate conclusions. The adaptively weighted Fisher's method (AW-Fisher), initially developed for omics applications but applicable for general meta-analysis, is an effective approach to combine P-values from K independent studies and to provide better biological interpretability by characterizing which studies contribute to the meta-analysis. Currently, AW-Fisher suffers from the lack of fast P-value computation and variability estimate of AW weights. When the number of studies K is large, the 3K - 1 possible differential expression pattern categories generated by AW-Fisher can become intractable. In this paper, we develop an importance sampling scheme with spline interpolation to increase the accuracy and speed of the P-value calculation. We also apply bootstrapping to construct a variability index for the AW-Fisher weight estimator and a co-membership matrix to categorize (cluster) differentially expressed genes based on their meta-patterns for intuitive biological investigations. RESULTS: The superior performance of the proposed methods is shown in simulations as well as two real omics meta-analysis applications to demonstrate its insightful biological findings. AVAILABILITY AND IMPLEMENTATION: An R package AWFisher (calling C++) is available at Bioconductor and GitHub (https://github.com/Caleb-Huo/AWFisher), and all datasets and programing codes for this paper are available in the Supplementary Material. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
Subject(s)
Genomics , Research Design , Biomarkers , GenomeABSTRACT
BACKGROUND: To compare the sensitivity and specificity of the recommended 2-step rapid antigen detection test (RADT) with confirmatory culture vs the point-of-care (POC) polymerase chain reaction (PCR) Roche cobas® Liat® Strep A test for detection of group A Streptococcus (GAS) in pediatric patients with pharyngitis, and to investigate the impact of these tests on antibiotic use in a large pediatric clinic. METHODS: This prospective, open-label study was conducted at a single site during fall/winter 2016-2017. A total of 275 patients aged 3 to 18 years with symptoms of pharyngitis had a throat-swab specimen analyzed using RADT, POC PCR, and culture. The sensitivity, specificity, and percentage agreement (95% CI) between assays and a laboratory-based nucleic acid amplification test were calculated. DNA sequencing was used to adjudicate discrepancies. The RADT or POC PCR result was provided to clinicians on alternating weeks to compare the impact on antibiotic use. RESULTS: A total of 255 samples were evaluated; 110 (43.1%) were GAS positive. Sensitivities (95% CI) for POC PCR, RADT, and culture were 95.5% (89.7-98.5%), 85.5% (77.5-1.5%), and 71.8% (62.4-80.0%), respectively. Specificities (95% CI) for POC PCR, RADT, and culture were 99.3% (96.2-99.98%), 93.7% (88.5-97.1%), and 100% (97.5-100%), respectively. Compared with RADT, POC PCR resulted in significantly greater appropriate antibiotic use (97.1% vs 87.5%; P = .0065). CONCLUSION: Under real-world conditions, RADT results were less specific and culture results were less sensitive than found in established literature and led to increased rates of inappropriate antibiotic use. POC PCR had high sensitivity and specificity and rapid turnaround times, and led to more appropriate antibiotic use. TRIAL REGISTRATION: ID number ISRCTN84562679 . Registered October 162,018, retrospectively registered.
Subject(s)
Anti-Bacterial Agents/therapeutic use , Pharyngitis/diagnosis , Pharyngitis/drug therapy , Point-of-Care Systems , Polymerase Chain Reaction , Streptococcal Infections/diagnosis , Streptococcal Infections/drug therapy , Streptococcus pyogenes , Adolescent , Child , Child, Preschool , Drug Utilization/statistics & numerical data , Humans , Pharyngitis/microbiology , Primary Health Care , Prospective Studies , Sensitivity and Specificity , Streptococcal Infections/microbiology , Streptococcus pyogenes/isolation & purificationABSTRACT
A main challenge in molecular diagnostic research is to accurately evaluate the performance of a new nucleic acid amplification test when the reference standard is imperfect. Several approaches, such as discrepant analysis, composite reference standard (CRS) method, or latent class analysis (LCA), are commonly applied for this purpose by combining multiple imperfect (reference) test results. In discrepant analysis or LCA, test results from the new assay are often involved in the construction of a new pseudo-reference standard, which results in the potential risk of overestimating the parameters of interest. On the contrary, the CRS methods only combine the results of reference tests, which is more preferable in practice. In this article, we study the properties of two extreme CRS methods, i.e., combining multiple reference test results by the "any positive" rule or by the "all-positive" rule, and propose a new approach "dual composite reference standards (dCRS)" based on these two extreme methods to reduce the biases of the estimates. Simulations are performed for various scenarios and the proposed approach is applied to two real datasets. The results demonstrate that our approach outperforms other commonly used approaches and therefore is recommended for future applications.
Subject(s)
Biomedical Research/statistics & numerical data , Diagnostic Tests, Routine/statistics & numerical data , Latent Class Analysis , Pathology, Molecular/statistics & numerical data , Reference Standards , Biomedical Research/methods , Chlamydia Infections/diagnosis , Clostridium Infections/diagnosis , Diagnostic Tests, Routine/methods , Humans , Pathology, Molecular/methodsABSTRACT
MOTIVATION: Supervised machine learning is commonly applied in genomic research to construct a classifier from the training data that is generalizable to predict independent testing data. When test datasets are not available, cross-validation is commonly used to estimate the error rate. Many machine learning methods are available, and it is well known that no universally best method exists in general. It has been a common practice to apply many machine learning methods and report the method that produces the smallest cross-validation error rate. Theoretically, such a procedure produces a selection bias. Consequently, many clinical studies with moderate sample sizes (e.g. n = 30-60) risk reporting a falsely small cross-validation error rate that could not be validated later in independent cohorts. RESULTS: In this article, we illustrated the probabilistic framework of the problem and explored the statistical and asymptotic properties. We proposed a new bias correction method based on learning curve fitting by inverse power law (IPL) and compared it with three existing methods: nested cross-validation, weighted mean correction and Tibshirani-Tibshirani procedure. All methods were compared in simulation datasets, five moderate size real datasets and two large breast cancer datasets. The result showed that IPL outperforms the other methods in bias correction with smaller variance, and it has an additional advantage to extrapolate error estimates for larger sample sizes, a practical feature to recommend whether more samples should be recruited to improve the classifier and accuracy. An R package 'MLbias' and all source files are publicly available. AVAILABILITY AND IMPLEMENTATION: tsenglab.biostat.pitt.edu/software.htm. CONTACT: ctseng@pitt.edu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
Subject(s)
Artificial Intelligence , Genomics/methods , Breast Neoplasms/genetics , Data Interpretation, Statistical , Female , Gene Expression Profiling , Humans , Models, Theoretical , Sample SizeABSTRACT
SUMMARY: With the rapid advances and prevalence of high-throughput genomic technologies, integrating information of multiple relevant genomic studies has brought new challenges. Microarray meta-analysis has become a frequently used tool in biomedical research. Little effort, however, has been made to develop a systematic pipeline and user-friendly software. In this article, we present MetaOmics, a suite of three R packages MetaQC, MetaDE and MetaPath, for quality control, differentially expressed gene identification and enriched pathway detection for microarray meta-analysis. MetaQC provides a quantitative and objective tool to assist study inclusion/exclusion criteria for meta-analysis. MetaDE and MetaPath were developed for candidate marker and pathway detection, which provide choices of marker detection, meta-analysis and pathway analysis methods. The system allows flexible input of experimental data, clinical outcome (case-control, multi-class, continuous or survival) and pathway databases. It allows missing values in experimental data and utilizes multi-core parallel computing for fast implementation. It generates informative summary output and visualization plots, operates on different operation systems and can be expanded to include new algorithms or combine different types of genomic data. This software suite provides a comprehensive tool to conveniently implement and compare various genomic meta-analysis pipelines. AVAILABILITY: http://www.biostat.pitt.edu/bioinfo/software.htm CONTACT: ctseng@pitt.edu SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.
Subject(s)
Gene Expression Profiling/methods , Genomics/methods , Microarray Analysis/methods , Software , Algorithms , Computational Biology/methods , Humans , Male , Meta-Analysis as Topic , Prostatic Neoplasms/genetics , Quality ControlABSTRACT
Trunk-like continuum robots have wide applications in manipulation and locomotion. In particular, trunk-like soft arms exhibit high dexterity and adaptability very similar to the creatures of the natural world. However, owing to the continuum and soft bodies, their performance in payload and spatial movements is limited. In this paper, we investigate the influence of key design parameters on robotic performance. It is verified that a larger workspace, lateral stiffness, payload, and bending moment could be achieved with adjustments to soft materials' hardness, the height of module segments, and arrayed radius of actuators. Especially, a 55% increase in arrayed radius would enhance the lateral stiffness by 25% and a bending moment by 55%. An 80% increase in segment height would enlarge 112% of the elongation range and 70 % of the bending range. Around 200% and 150% increments in the segment's lateral stiffness and payload forces, respectively, could be obtained by tuning the hardness of soft materials. These relations enable the design customization of trunk-like soft arms, in which this tapering structure ensures stability via the stocky base for an impact reduction of 50% compared to that of the tip and ensures dexterity of the long tip for a relatively larger bending range of over 400% compared to that of the base. The complete methodology of the design concept, analytical models, simulation, and experiments is developed to offer comprehensive guidelines for trunk-like soft robotic design and enable high performance in robotic manipulation.
ABSTRACT
The median failure time is often utilized to summarize survival data because it has a more straightforward interpretation for investigators in practice than the popular hazard function. However, existing methods for comparing median failure times for censored survival data either require estimation of the probability density function or involve complicated formulas to calculate the variance of the estimates. In this article, we modify a K-sample median test for censored survival data (Brookmeyer and Crowley, 1982, Journal of the American Statistical Association 77, 433-440) through a simple contingency table approach where each cell counts the number of observations in each sample that are greater than the pooled median or vice versa. Under censoring, this approach would generate noninteger entries for the cells in the contingency table. We propose to construct a weighted asymptotic test statistic that aggregates dependent χ(2)-statistics formed at the nearest integer points to the original noninteger entries. We show that this statistic follows approximately a χ(2)-distribution with k-1 degrees of freedom. For a small sample case, we propose a test statistic based on combined p-values from Fisher's exact tests, which follows a χ(2)-distribution with 2 degrees of freedom. Simulation studies are performed to show that the proposed method provides reasonable type I error probabilities and powers. The proposed method is illustrated with two real datasets from phase III breast cancer clinical trials.
Subject(s)
Models, Statistical , Survival Analysis , Biometry , Breast Neoplasms/mortality , Breast Neoplasms/therapy , Combined Modality Therapy , Data Interpretation, Statistical , Databases, Factual/statistics & numerical data , Female , Humans , Kaplan-Meier Estimate , Proportional Hazards Models , Randomized Controlled Trials as Topic/statistics & numerical dataABSTRACT
Background: Rapid antigen detection tests (RADTs) are the standard of care (SOC) for testing in patients with suspected group A β-hemolytic Streptococcus (Strep A) infection. Due to lower sensitivity, guidelines recommend confirmatory microbiological culture following negative RADT results. This process is time-consuming, and adherence is often poor, resulting in high rates of inappropriate antibiotic prescribing. We sought to evaluate the impact of switching from RADTs to point-of-care (POC) polymerase chain reaction (PCR) testing on use of antibiotics in primary care, when used as part of an antibiotic stewardship initiative. Methods: In this retrospective before-after study, electronic medical records of any patients presenting with suspected acute pharyngitis (June 2018-May 2019) across 15 outpatient primary care clinics were evaluated. Strep A was detected using the cobas Strep A assay (cobas Liat system). Results: Analysis of 10 081 eligible patient records showed that POC PCR testing resulted in a 44.1% reduction in antibiotic prescribing for patients with a negative POC PCR test result (10.1% PCR vs 18.0% RADT; P < .0001). Rates of antibiotic prescription varied across clinical sites, ranging between 10.7% and 33.8% and 12.4% and 34.4% during the use of PCR tests and RADTs, respectively. POC PCR had no impact on prescription rates in patients with positive POC test results compared to RADTs (76.2% vs 76.5%, respectively). More than 99% of antibiotics were prescribed during the initial primary care encounter. Conclusions: As part of a broader antibiotic stewardship initiative, implementation of POC PCR as SOC in outpatients with acute pharyngitis symptoms reduced the volume of inappropriate antibiotic prescriptions.
ABSTRACT
AIMS: Point-of-care (POC) tests for influenza and respiratory syncytial virus (RSV) offer the potential to improve patient management and antimicrobial stewardship. Studies have focused on performance; however, no workflow assessments have been published comparing POC molecular tests. This study compared the Liat and ID Now systems workflow, to assist end-users in selecting an influenza and/or RSV POC test. METHODS: Staffing, walk-away and turnaround time (TAT) of the Liat and ID Now systems were determined using 40 nasopharyngeal samples, positive for influenza or RSV. The ID Now system requires separate tests for influenza and RSV, so parallel (two instruments) and sequential (one instrument) workflows were evaluated. RESULTS: The ID Now ranged 4.1-6.2 min for staffing, 1.9-10.9 min for walk-away and 6.4-15.8 min for TAT per result. The Liat ranged 1.1-1.8 min for staffing, 20.0-20.5 min for walk-away and 21.3-22.0 min for TAT. Mean walk-away time comprised 38.0% (influenza positive) and 68.1% (influenza negative) of TAT for ID Now and 93.7% (influenza/RSV) for Liat. The ID Now parallel workflow resulted in medians of 5.9 min for staffing, 9.7 min for walk-away and 15.6 min for TAT. Assuming prevalence of 20% influenza and 20% RSV, the ID Now sequential workflow resulted in medians of 9.4 min for staffing, 17.4 min for walk-away, and 27.1 min for TAT. CONCLUSIONS: The ID Now and Liat systems offer different workflow characteristics. Key considerations for implementation include value of both influenza and RSV results, clinical setting, staffing capacity, and instrument(s) placement.
Subject(s)
Influenza A virus/isolation & purification , Influenza B virus/isolation & purification , Influenza, Human/diagnosis , Point-of-Care Testing , Respiratory Syncytial Virus Infections/diagnosis , Respiratory Syncytial Virus, Human/isolation & purification , Humans , Influenza, Human/virology , Nasopharynx/virology , Respiratory Syncytial Virus Infections/virology , WorkflowABSTRACT
Clostridium difficile infection (CDI) is a high burden and significant cause of healthcare-acquired infectious diarrhea in the United States (US). Timely and accurate diagnosis of CDI enables the rapid initiation of antibiotic therapy and infection control policies to minimize disease transmission. Polymerase chain reaction (PCR) assays have become a preferred modality for diagnosing CDI in the US. The cobas Liat Cdiff PCR test is a novel assay that can be performed on-demand for hospital-based testing with a rapid 20-minute turnaround time from specimen collection to result reporting. We compared the clinical performance of the cobas Liat Cdiff test to the previously introduced Xpert C. difficile/Epi test; both tests are FDA-cleared PCR assays that detect the toxin B (tcdB) gene of C. difficile. Prospectively collected and remnant stool specimens from 310 patients with suspected CDI were obtained for analysis. The cobas Liat Cdiff and Xpert PCR tests showed an overall percent agreement of 97.4% (302/310; 95% CI: 95.0-98.9). Low bacterial burdens of toxigenic C. difficile, indicated by significantly delayed PCR cycle threshold (Ct) values, explained most of the discordance. Positive and negative percent agreement of the cobas Liat Cdiff test compared to the Xpert PCR test were 94.5% (52/55) and 98.0% (250/255), respectively. The clinical performance of the cobas Liat Cdiff test, combined with its simplicity of use and rapid result reporting, provides a reliable option for clinical laboratory use.
Subject(s)
Clostridioides difficile/isolation & purification , Clostridium Infections/diagnosis , Diagnostic Tests, Routine , Feces/microbiology , Molecular Diagnostic Techniques , Bacterial Proteins/genetics , Bacterial Toxins/genetics , Clostridioides difficile/genetics , Enterotoxins/genetics , Humans , Polymerase Chain Reaction , Prospective Studies , Sensitivity and SpecificityABSTRACT
Microarray analysis to monitor expression activities in thousands of genes simultaneously has become routine in biomedical research during the past decade. a tremendous amount of expression profiles are generated and stored in the public domain and information integration by meta-analysis to detect differentially expressed (DE) genes has become popular to obtain increased statistical power and validated findings. Methods that aggregate transformed p-value evidence have been widely used in genomic settings, among which Fisher's and Stouffer's methods are the most popular ones. In practice, raw data and p-values of DE evidence are often not available in genomic studies that are to be combined. Instead, only the detected DE gene lists under a certain p-value threshold (e.g., DE genes with p-value < 0.001) are reported in journal publications. The truncated p-value information makes the aforementioned meta-analysis methods inapplicable and researchers are forced to apply a less efficient vote counting method or naïvely drop the studies with incomplete information. The purpose of this paper is to develop effective meta-analysis methods for such situations with partially censored p-values. We developed and compared three imputation methods-mean imputation, single random imputation and multiple imputation-for a general class of evidence aggregation methods of which Fisher's and Stouffer's methods are special examples. The null distribution of each method was analytically derived and subsequent inference and genomic analysis frameworks were established. Simulations were performed to investigate the type Ierror, power and the control of false discovery rate (FDR) for (correlated) gene expression data. The proposed methods were applied to several genomic applications in colorectal cancer, pain and liquid association analysis of major depressive disorder (MDD). The results showed that imputation methods outperformed existing naïve approaches. Mean imputation and multiple imputation methods performed the best and are recommended for future applications.