False-positive Threshold Research Articles

Efficient False Positive Control Algorithms in Big Data Mining

The typical hypothesis testing issue in statistical analysis is determining whether a pattern is significantly associated with a specific class label. This usually leads to highly challenging multiple-hypothesis testing problems in big data mining scenarios, as millions or billions of hypothesis tests in large-scale exploratory data analysis can result in a large number of false positive results. The permutation testing-based FWER control method (PFWER) is theoretically effective in dealing with multiple hypothesis testing issues. In reality, however, this theoretical approach confronts a serious computational efficiency problem. It takes an extremely long time to compute an appropriate FWER false positive control threshold using PFWER, which is almost impossible to achieve in a reasonable amount of time using human effort on medium- or large-scale data. Although some methods for improving the efficiency of the FWER false positive control threshold calculation have been proposed, most of them are stand-alone, and there is still a lot of space for efficiency improvement. To address this problem, this paper proposes a distributed PFWER false-positive threshold calculation method for large-scale data. The computational effectiveness increases significantly when compared to the current approaches. The FP-growth algorithm is used first for pattern mining, and the mining process reduces the computation of invalid patterns by using pruning operations and index optimization for merging patterns with index transactions. The distributed computing technique is introduced on this basis, and the constructed FP tree is decomposed into a set of subtrees, each corresponding to a subtask. All subtrees (subtasks) are distributed to different computing nodes. Each node independently calculates the local significance threshold according to the designated subtasks. Finally, all local results are aggregated to compute the FWER false positive control threshold, which is completely consistent with the theoretical result. A series of experimental findings on 11 real-world datasets demonstrate that the distributed algorithm proposed in this paper can significantly improve the computation efficiency of PFWER while ensuring its theoretical accuracy.

Three-dimensional reconstruction and virtual reposition of fragments compared to two dimensional measurements of midshaft clavicle fracture shortening

BackgroundMidshaft clavicle fracture shortening measurement is a reported key element for indication to surgical management and reporting of clinical trials. Determination of pre-fracture clavicle length for shortening measurement remains an unresolved issue. The purpose of the study was to assess accuracy of a novel technique of three-dimensional reconstruction and virtual reposition of bone fragments (3D-VR) for determination of pre-fracture clavicle length and measurement of shortening.MethodsAccuracy of 3D-VR measurements was assessed using 5 synthetic bone clavicle fracture models. Measurements were compared between caliper and 3D-VR technique measurements. Correlation between 3D-VR and 2D measurements on standard radiographs was assessed on a cohort of 20 midshaft fractures. Four different methods for 2D measurements were assessed.ResultsMean difference between caliper measurements and 3D-VR was 0.74 mm (95CI = − 2.51;3.98) (p = 0.56) on synthetic fracture models. Mean differences between 3D-VR and standard radiograph shortening measurement methods were 11.95 mm (95CI = 7.44;16.46) for method 1 (Jeray et al.) and 9.28 mm (95CI = 4.77;13.79) for method 2 (Smekal et al.) (p < 0.05). Differences were − 1.02 mm (95CI = − 5.53;3.48) for method 3 (Silva et al.) and − 2.04 mm (95CI = − 6.55;2.47) for method 4 (own method). Interobserver correlation ranged between 0.85 and 0.99. A false positive threshold of 20 mm was measured by the two observers in 25% of the case according to method of method 1, 30–35% with method 2, 15% with method 3 et al. and 5–10% with the method 4.Conclusion3D VR is accurate in measuring midshaft clavicle fracture length and shortening. Two dimensional measurements may be used for approximation of clavicular shortening.

A-12 Racial Differences in Performance Validity Test Failure Rates

Abstract Objective Embedded performance validity tests (PVTs) may show increased false positive rates in racially diverse examinees. This study examined false positive rates by race in an older adult sample. Method The project involved secondary analysis of a deidentified dataset (N = 22,688) from the National Alzheimer’s Coordinating Center (NACC). Participants were included if their identified race was African American or white. Exclusion criteria included diagnosis of mild cognitive impairment (MCI; n = 5160) or dementia (n = 5550). The initial sample included 11,114 participants grouped as cognitively normal (89.9%) or impaired but not MCI of whom 16.4% identified as African American. Propensity score matching was conducted by diagnostic group to match African American and white participants on age, education, gender, and MMSE score. The final sample included 3024 and 482 participants in normal and impaired groups, respectively, with 50% of participants identifying as African American in each group. Failure rates on five embedded PVTs in the NACC cognitive test battery were examined by race and by diagnosis. Results Age, education, gender, and MMSE score were not significantly different by race in either diagnostic group. In the normal group, 4.7% of African American and 1.9% of white participants failed two or more PVTs (p &amp;lt; 0.001). In the impaired group, 9.5% of African American and 5.8% of white participants failed two or more PVTs (n.s.). Conclusions PVT failure rates were significantly higher among African American participants in the normal group but not in the impaired group. Failure rates remained below a common false positive threshold of 10%.

A-184 Performance Validity Test Failure Rates in Bilingual Individuals Evaluated in English

Abstract Objective Classification accuracy of embedded performance validity tests (PVTs) is unknown in cases involving bilingual examinees evaluated in English. This study examined false positive rates in bilingual individuals in an older adult sample. Method The project involved secondary analysis of a deidentified dataset (N = 22,688) from the National Alzheimer’s Coordinating Center (NACC). Exclusion criteria were diagnosis of mild cognitive impairment (MCI; n = 5160) or dementia (n = 5550). The initial sample included 11,513 participants grouped as cognitively normal (89.6%) or impaired but not MCI. A subset of 275 participants was identified with a primary language other than English who were evaluated in English. Propensity score matching was conducted by diagnostic group to match bilingual to monolingual participants on age, education, gender, and MMSE score. The final sample included 450 and 100 participants in normal and impaired groups, respectively. Failure rates on five embedded PVTs in the NACC cognitive test battery were examined by language and by diagnosis. Results Age, education, gender, and MMSE score were not significantly different by language in either diagnostic group. In the normal group, 4.9% of bilingual and 2.2% of monolingual participants failed two or more PVTs (n.s.). In the impaired group, 12% of bilingual and 6% of monolingual participants failed two or more PVTs (n.s.). Conclusions PVT failure rates were not significantly different between bilingual participants evaluated in English and monolingual participants in either diagnostic group. Failure rates, however, increased slightly above a common false positive threshold of 10% in bilingual participants in the impaired group.

Associations of annual ambient PM2.5 components with DNAm PhenoAge acceleration in elderly men: The Normative Aging Study

Current studies indicate that long-term exposure to ambient fine particulate matter (PM2.5) is related with global mortality, yet no studies have explored relationships of PM2.5 and its species with DNAm PhenoAge acceleration (DNAmPhenoAccel), a new epigenetic biomarker of phenotypic age. We identified which PM2.5 species had association with DNAmPhenoAccel in a one-year exposure window in a longitudinal cohort. We collected whole blood samples from 683 elderly men in the Normative Aging Study between 1999 and 2013 (n = 1254 visits). DNAm PhenoAge was calculated using 513 CpGs retrieved from the Illumina Infinium HumanMethylation450 BeadChip. Daily concentrations of PM2.5 species were measured at a fixed air-quality monitoring site and one-year moving averages were computed. Linear mixed-effect (LME) regression and Bayesian kernel machine (BKM) regression were used to estimate the associations. The covariates included chronological age, body mass index (BMI), cigarette pack years, smoking status, estimated cell types, batch effects etc. Benjamini-Hochberg false discovery rate at a 5% false positive threshold was used to adjust for multiple comparison. During the study period, the mean DNAm PhenoAge and chronological age in our subjects were 68 and 73 years old, respectively. Using LME model, only lead and calcium were significantly associated with DNAmPhenoAccel. For example, an interquartile range (IQR, 0.0011 μg/m3) increase in lead was associated with a 1.29-year [95% confidence interval (CI): 0.47, 2.11] increase in DNAmPhenoAccel. Using BKM model, we selected PM2.5, lead, and silicon to be predictors for DNAmPhenoAccel. A subsequent LME model showed that only lead had significant effect on DNAmPhenoAccel: 1.45-year (95% CI: 0.46, 2.46) increase in DNAmPhenoAccel following an IQR increase in one-year lead. This is the first study that investigates long-term effects of PM2.5 components on DNAmPhenoAccel. The results demonstrate that lead and calcium contained in PM2.5 was robustly associated with DNAmPhenoAccel.

Assessing the Accuracy of Variant Detection in Cost-Effective Gene Panel Testing by Next-Generation Sequencing

There is significant debate within the diagnostics community regarding the accuracy of variant identification by next-generation sequencing and the necessity of confirmatory testing of detected variants. Because the quality threshold to discriminate false positives depends on the workflow, no regulatory standard regarding this matter has yet been published. The goal of this study was to empirically determine the threshold to perform additional Sanger sequencing and to reduce the experimental cost to a practical level. Using 278 model genes, a hybridization capture-based protocol was examined to meet the clinical requirements of low cost, high efficiency, and high-quality data. To reduce excessive false-positive detection, filtering processes were introduced to remove mismapped reads and strand-biased detection to a published best-practices pipeline. With seven samples from the 1000 Genomes Project, 2750 single-nucleotide polymorphisms and 142 insertions/deletions were identified by our designed workflow. Compared with variants registered in the single nucleotide polymorphism database (dbSNP), a zero false-positive threshold value was determined (quality score>1000). The variants satisfying these criteria accounted for 95.6% of single-nucleotide polymorphisms and 50.7% of insertions/deletions. Except for deletions located within the highly repeated sequences, the workflow achieved 100% sensitivity. The established threshold allowed us to discriminate between convincing variants and those requiring validation, a design that reconciles the competing objectives of cost minimization and quality maximization of clinical gene panel testing.

Detection of and Compensation for EMG Disturbances for Powered Lower Limb Prosthesis Control.

Myoelectric pattern recognition algorithms have been proposed for the control of powered lower limb prostheses, but electromyography (EMG) signal disturbances remain an obstacle to clinical implementation. To address this problem, we used a log-likelihood metric to detect simulated EMG disturbances and real disturbances acquired from EMG containing electrode shift. We found that features extracted from disturbed EMG have much lower log likelihoods than those from undisturbed signals and can be detected using a single threshold acquired from the training data. We designed a linear discriminant analysis (LDA) classifier that uses the log likelihood to decide between using a combination of EMG and mechanical sensors and using mechanical sensors only, to predict locomotion modes. When EMG contained disturbances, our classifier detected those disturbances and disregarded EMG data. Our classifier had significantly lower errors than a standard LDA classifier in the presence of EMG disturbances. The log-likelihood classifier had a low false positive threshold, and thus did not perform significantly differently from the standard LDA classifier when EMG did not contain disturbances. The log-likelihood threshold could also be applied to individual EMG channels, enabling specific channels containing EMG disturbances to be appropriately ignored when making locomotion mode predictions.

Applying label-free quantitation to top down proteomics.

With the prospect of resolving whole protein molecules into their myriad proteoforms on a proteomic scale, the question of their quantitative analysis in discovery mode comes to the fore. Here, we demonstrate a robust pipeline for the identification and stringent scoring of abundance changes of whole protein forms <30 kDa in a complex system. The input is ∼100–400 μg of total protein for each biological replicate, and the outputs are graphical displays depicting statistical confidence metrics for each proteoform (i.e., a volcano plot and representations of the technical and biological variation). A key part of the pipeline is the hierarchical linear model that is tailored to the original design of the study. Here, we apply this new pipeline to measure the proteoform-level effects of deleting a histone deacetylase (rpd3) in S. cerevisiae. Over 100 proteoform changes were detected above a 5% false positive threshold in WT vs the Δrpd3 mutant, including the validating observation of hyperacetylation of histone H4 and both H2B isoforms. Ultimately, this approach to label-free top down proteomics in discovery mode is a critical technical advance for testing the hypothesis that whole proteoforms can link more tightly to complex phenotypes in cell and disease biology than do peptides created in shotgun proteomics.