Browsing by Subject "Bayes theorem"

Now showing 1 - 8 of 8
  • Thumbnail Image
    Item
    circMeta: a unified computational framework for genomic feature annotation and differential expression analysis of circular RNAs
    (Oxford University Press, 2020-01-15) Chen, Li; Wang, Feng; Bruggeman, Emily C.; Li, Chao; Yao, Bing; Medicine, School of Medicine
    Motivation: Circular RNAs (circRNAs), a class of non-coding RNAs generated from non-canonical back-splicing events, have emerged to play key roles in many biological processes. Though numerous tools have been developed to detect circRNAs from rRNA-depleted RNA-seq data based on back-splicing junction-spanning reads, computational tools to identify critical genomic features regulating circRNA biogenesis are still lacking. In addition, rigorous statistical methods to perform differential expression (DE) analysis of circRNAs remain under-developed. Results: We present circMeta, a unified computational framework for circRNA analyses. circMeta has three primary functional modules: (i) a pipeline for comprehensive genomic feature annotation related to circRNA biogenesis, including length of introns flanking circularized exons, repetitive elements such as Alu elements and SINEs, competition score for forming circulation and RNA editing in back-splicing flanking introns; (ii) a two-stage DE approach of circRNAs based on circular junction reads to quantitatively compare circRNA levels and (iii) a Bayesian hierarchical model for DE analysis of circRNAs based on the ratio of circular reads to linear reads in back-splicing sites to study spatial and temporal regulation of circRNA production. Both proposed DE methods without and with considering host genes outperform existing methods by obtaining better control of false discovery rate and comparable statistical power. Moreover, the identified DE circRNAs by the proposed two-stage DE approach display potential biological functions in Gene Ontology and circRNA-miRNA-mRNA networks that are not able to be detected using existing mRNA DE methods. Furthermore, top DE circRNAs have been further validated by RT-qPCR using divergent primers spanning back-splicing junctions.
  • Thumbnail Image
    Item
    Effect of Therapeutic Hypothermia Initiated After 6 Hours of Age on Death or Disability Among Newborns With Hypoxic-Ischemic Encephalopathy: A Randomized Clinical Trial
    (American Medical Association, 2017-10-24) Laptook, Abbot R.; Shankaran, Seetha; Tyson, Jon E.; Munoz, Breda; Bell, Edward F.; Goldberg, Ronald N.; Parikh, Nehal A.; Ambalavanan, Namasivayam; Pedroza, Claudia; Pappas, Athina; Das, Abhik; Chaudhary, Aasma S.; Ehrenkranz, Richard A.; Hensman, Angelita M.; Van Meurs, Krisa P.; Chalak, Lina F.; Hamrick, Shannon E. G.; Sokol, Gregory M.; Walsh, Michele C.; Poindexter, Brenda B.; Faix, Roger G.; Watterberg, Kristi L.; Frantz, Ivan D., III; Guillet, Ronnie; Devaskar, Uday; Truog, William E.; Chock, Valerie Y.; Wyckoff, Myra H.; McGowan, Elisabeth C.; Carlton, David P.; Harmon, Heidi M.; Brumbaugh, Jane E.; Cotten, C. Michael; Sánchez, Pablo J.; Hibbs, Anna Maria; Higgins, Rosemary D.; Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network; Pediatrics, School of Medicine
    Importance: Hypothermia initiated at less than 6 hours after birth reduces death or disability for infants with hypoxic-ischemic encephalopathy at 36 weeks' or later gestation. To our knowledge, hypothermia trials have not been performed in infants presenting after 6 hours. Objective: To estimate the probability that hypothermia initiated at 6 to 24 hours after birth reduces the risk of death or disability at 18 months among infants with hypoxic-ischemic encephalopathy. Design, Setting, and Participants: A randomized clinical trial was conducted between April 2008 and June 2016 among infants at 36 weeks' or later gestation with moderate or severe hypoxic-ischemic encephalopathy enrolled at 6 to 24 hours after birth. Twenty-one US Neonatal Research Network centers participated. Bayesian analyses were prespecified given the anticipated limited sample size. Interventions: Targeted esophageal temperature was used in 168 infants. Eighty-three hypothermic infants were maintained at 33.5°C (acceptable range, 33°C-34°C) for 96 hours and then rewarmed. Eighty-five noncooled infants were maintained at 37.0°C (acceptable range, 36.5°C-37.3°C). Main Outcomes and Measures: The composite of death or disability (moderate or severe) at 18 to 22 months adjusted for level of encephalopathy and age at randomization. Results: Hypothermic and noncooled infants were term (mean [SD], 39 [2] and 39 [1] weeks' gestation, respectively), and 47 of 83 (57%) and 55 of 85 (65%) were male, respectively. Both groups were acidemic at birth, predominantly transferred to the treating center with moderate encephalopathy, and were randomized at a mean (SD) of 16 (5) and 15 (5) hours for hypothermic and noncooled groups, respectively. The primary outcome occurred in 19 of 78 hypothermic infants (24.4%) and 22 of 79 noncooled infants (27.9%) (absolute difference, 3.5%; 95% CI, -1% to 17%). Bayesian analysis using a neutral prior indicated a 76% posterior probability of reduced death or disability with hypothermia relative to the noncooled group (adjusted posterior risk ratio, 0.86; 95% credible interval, 0.58-1.29). The probability that death or disability in cooled infants was at least 1%, 2%, or 3% less than noncooled infants was 71%, 64%, and 56%, respectively. Conclusions and Relevance: Among term infants with hypoxic-ischemic encephalopathy, hypothermia initiated at 6 to 24 hours after birth compared with noncooling resulted in a 76% probability of any reduction in death or disability, and a 64% probability of at least 2% less death or disability at 18 to 22 months. Hypothermia initiated at 6 to 24 hours after birth may have benefit but there is uncertainty in its effectiveness.
  • Thumbnail Image
    Item
    Evaluation of Plasma Phosphorylated Tau217 for Differentiation Between Alzheimer Disease and Frontotemporal Lobar Degeneration Subtypes Among Patients With Corticobasal Syndrome
    (American Medical Association, 2023) VandeVrede, Lawren; La Joie, Renaud; Thijssen, Elisabeth H.; Asken, Breton M.; Vento, Stephanie A.; Tsuei, Torie; Baker, Suzanne L.; Cobigo, Yann; Fonseca, Corrina; Heuer, Hilary W.; Kramer, Joel H.; Ljubenkov, Peter A.; Rabinovici, Gil D.; Rojas, Julio C.; Rosen, Howie J.; Staffaroni, Adam M.; Boeve, Brad F.; Dickerson, Brad C.; Grossman, Murray; Huey, Edward D.; Irwin, David J.; Litvan, Irene; Pantelyat, Alexander Y.; Tartaglia, Maria Carmela; Dage, Jeffrey L.; Boxer, Adam L.; Neurology, School of Medicine
    Importance: Plasma phosphorylated tau217 (p-tau217), a biomarker of Alzheimer disease (AD), is of special interest in corticobasal syndrome (CBS) because autopsy studies have revealed AD is the driving neuropathology in up to 40% of cases. This differentiates CBS from other 4-repeat tauopathy (4RT)-associated syndromes, such as progressive supranuclear palsy Richardson syndrome (PSP-RS) and nonfluent primary progressive aphasia (nfvPPA), where underlying frontotemporal lobar degeneration (FTLD) is typically the primary neuropathology. Objective: To validate plasma p-tau217 against positron emission tomography (PET) in 4RT-associated syndromes, especially CBS. Design, setting, and participants: This multicohort study with 6, 12, and 24-month follow-up recruited adult participants between January 2011 and September 2020 from 8 tertiary care centers in the 4RT Neuroimaging Initiative (4RTNI). All participants with CBS (n = 113), PSP-RS (n = 121), and nfvPPA (n = 39) were included; other diagnoses were excluded due to rarity (n = 29). Individuals with PET-confirmed AD (n = 54) and PET-negative cognitively normal control individuals (n = 59) were evaluated at University of California San Francisco. Operators were blinded to the cohort. Main outcome and measures: Plasma p-tau217, measured by Meso Scale Discovery electrochemiluminescence, was validated against amyloid-β (Aβ) and flortaucipir (FTP) PET. Imaging analyses used voxel-based morphometry and bayesian linear mixed-effects modeling. Clinical biomarker associations were evaluated using longitudinal mixed-effect modeling. Results: Of 386 participants, 199 (52%) were female, and the mean (SD) age was 68 (8) years. Plasma p-tau217 was elevated in patients with CBS with positive Aβ PET results (mean [SD], 0.57 [0.43] pg/mL) or FTP PET (mean [SD], 0.75 [0.30] pg/mL) to concentrations comparable to control individuals with AD (mean [SD], 0.72 [0.37]), whereas PSP-RS and nfvPPA showed no increase relative to control. Within CBS, p-tau217 had excellent diagnostic performance with area under the receiver operating characteristic curve (AUC) for Aβ PET of 0.87 (95% CI, 0.76-0.98; P < .001) and FTP PET of 0.93 (95% CI, 0.83-1.00; P < .001). At baseline, individuals with CBS-AD (n = 12), defined by a PET-validated plasma p-tau217 cutoff 0.25 pg/mL or greater, had increased temporoparietal atrophy at baseline compared to individuals with CBS-FTLD (n = 39), whereas longitudinally, individuals with CBS-FTLD had faster brainstem atrophy rates. Individuals with CBS-FTLD also progressed more rapidly on a modified version of the PSP Rating Scale than those with CBS-AD (mean [SD], 3.5 [0.5] vs 0.8 [0.8] points/year; P = .005). Conclusions and relevance: In this cohort study, plasma p-tau217 had excellent diagnostic performance for identifying Aβ or FTP PET positivity within CBS with likely underlying AD pathology. Plasma P-tau217 may be a useful and inexpensive biomarker to select patients for CBS clinical trials.
  • Thumbnail Image
    Item
    Improving the diagnosis of severe malaria in African children using platelet counts and plasma PfHRP2 concentrations
    (American Association for the Advancement of Science, 2022) Watson, James A.; Uyoga, Sophie; Wanjiku, Perpetual; Makale, Johnstone; Nyutu, Gideon M.; Mturi, Neema; George, Elizabeth C.; Woodrow, Charles J.; Day, Nicholas P. J.; Bejon, Philip; Opoka, Robert O.; Dondorp, Arjen M.; John, Chandy C.; Maitland, Kathryn; Williams, Thomas N.; White, Nicholas J.; Pediatrics, School of Medicine
    Severe malaria caused by Plasmodium falciparum is difficult to diagnose accurately in children in high-transmission settings. Using data from 2649 pediatric and adult patients enrolled in four studies of severe illness in three countries (Bangladesh, Kenya, and Uganda), we fitted Bayesian latent class models using two diagnostic markers: the platelet count and the plasma concentration of P. falciparum histidine-rich protein 2 (PfHRP2). In severely ill patients with clinical features consistent with severe malaria, the combination of a platelet count of ≤150,000/μl and a plasma PfHRP2 concentration of ≥1000 ng/ml had an estimated sensitivity of 74% and specificity of 93% in identifying severe falciparum malaria. Compared with misdiagnosed children, pediatric patients with true severe malaria had higher parasite densities, lower hematocrits, lower rates of invasive bacterial disease, and a lower prevalence of both sickle cell trait and sickle cell anemia. We estimate that one-third of the children enrolled into clinical studies of severe malaria in high-transmission settings in Africa had another cause of their severe illness.
  • Thumbnail Image
    Item
    Machine Learning Techniques for Prediction of Early Childhood Obesity
    (Schattauer, 2015-08-12) Dugan, T.M.; Mukhopadhyay, S.; Carroll, A.; Downs, S.; Department of Computer and Information Science, School of Science
    Objectives This paper aims to predict childhood obesity after age two, using only data collected prior to the second birthday by a clinical decision support system called CHICA. Methods Analyses of six different machine learning methods: RandomTree, RandomForest, J48, ID3, Naïve Bayes, and Bayes trained on CHICA data show that an accurate, sensitive model can be created. Results Of the methods analyzed, the ID3 model trained on the CHICA dataset proved the best overall performance with accuracy of 85% and sensitivity of 89%. Additionally, the ID3 model had a positive predictive value of 84% and a negative predictive value of 88%. The structure of the tree also gives insight into the strongest predictors of future obesity in children. Many of the strongest predictors seen in the ID3 modeling of the CHICA dataset have been independently validated in the literature as correlated with obesity, thereby supporting the validity of the model. Conclusions This study demonstrated that data from a production clinical decision support system can be used to build an accurate machine learning model to predict obesity in children after age two.
  • Thumbnail Image
    Item
    Performance of deep learning restoration methods for the extraction of particle dynamics in noisy microscopy image sequences
    (American Society for Cell Biology, 2021-04-19) Kefer, Paul; Iqbal, Fadil; Locatelli, Maelle; Lawrimore, Josh; Zhang, Mengdi; Bloom, Kerry; Bonin, Keith; Vidi, Pierre-Alexandre; Liu, Jing; Physics, School of Science
    Particle tracking in living systems requires low light exposure and short exposure times to avoid phototoxicity and photobleaching and to fully capture particle motion with high-speed imaging. Low-excitation light comes at the expense of tracking accuracy. Image restoration methods based on deep learning dramatically improve the signal-to-noise ratio in low-exposure data sets, qualitatively improving the images. However, it is not clear whether images generated by these methods yield accurate quantitative measurements such as diffusion parameters in (single) particle tracking experiments. Here, we evaluate the performance of two popular deep learning denoising software packages for particle tracking, using synthetic data sets and movies of diffusing chromatin as biological examples. With synthetic data, both supervised and unsupervised deep learning restored particle motions with high accuracy in two-dimensional data sets, whereas artifacts were introduced by the denoisers in three-dimensional data sets. Experimentally, we found that, while both supervised and unsupervised approaches improved tracking results compared with the original noisy images, supervised learning generally outperformed the unsupervised approach. We find that nicer-looking image sequences are not synonymous with more precise tracking results and highlight that deep learning algorithms can produce deceiving artifacts with extremely noisy images. Finally, we address the challenge of selecting parameters to train convolutional neural networks by implementing a frugal Bayesian optimizer that rapidly explores multidimensional parameter spaces, identifying networks yielding optimal particle tracking accuracy. Our study provides quantitative outcome measures of image restoration using deep learning. We anticipate broad application of this approach to critically evaluate artificial intelligence solutions for quantitative microscopy.
  • Thumbnail Image
    Item
    Risk Factors for Fracture in Patients with Coexisting Chronic Kidney Disease and Type 2 Diabetes: An Observational Analysis from the CREDENCE Trial
    (Hindawi, 2022-05-27) Young, Tamara K.; Toussaint, Nigel D.; Di Tanna, Gian Luca; Arnott, Clare; Hockham, Carinna; Kang, Amy; Schutte, Aletta E.; Perkovic, Vlado; Mahaffey, Kenneth W.; Agarwal, Rajiv; Bakris, George L.; Charytan, David M.; Heerspink, Hiddo J.L.; Levin, Adeera; Pollock, Carol; Wheeler, David C.; Zhang, Hong; Jardine, Meg J.; Medicine, School of Medicine
    Background: The fracture pathophysiology associated with type 2 diabetes and chronic kidney disease (CKD) is incompletely understood. We examined individual fracture predictors and prediction sets based on different pathophysiological hypotheses, testing whether any of the sets improved prediction beyond that based on traditional osteoporotic risk factors. Methods: Within the CREDENCE cohort with adjudicated fracture outcomes, we assessed the association of individual factors with fracture using Cox regression models. We used the Akaike information criteria (AIC) and Schwartz Bayes Criterion (SBC) to assess six separate variable sets based on hypothesized associations with fracture, namely, traditional osteoporosis, exploratory general population findings, cardiovascular risk, CKD-mineral and bone disorder, diabetic osteodystrophy, and an all-inclusive set containing all variables. Results: Fracture occurred in 135 (3.1%) participants over a median 2.35 [1.88-2.93] years. Independent fracture predictors were older age (hazard ratio [HR] 1.04, confidence interval [CI] 1.01-1.06), female sex (HR 2.49, CI 1.70-3.65), previous fracture (HR 2.30, CI 1.58-3.34), Asian race (HR 1.74, CI 1.09-2.78), vitamin D therapy requirement (HR 2.05, CI 1.31-3.21), HbA1c (HR 1.14, CI 1.00-1.32), prior cardiovascular event (HR 1.60, CI 1.10-2.33), and serum albumin (HR 0.41, CI 0.23-0.74) (lower albumin associated with greater risk). The goodness of fit of the various hypothesis sets was similar (AIC range 1870.92-1849.51, SBC range 1875.60-1948.04). Conclusion: Independent predictors of fracture were identified in the CREDENCE participants with type 2 diabetes and CKD. Fracture prediction was not improved by models built on alternative pathophysiology hypotheses compared with traditional osteoporosis predictors.
  • Thumbnail Image
    Item
    Symptom-Based COVID-19 Prognosis through AI-Based IoT: A Bioinformatics Approach
    (Hindawi, 2022-07-23) Pal, Madhumita; Parija, Smita; Mohapatra, Ranjan K.; Mishra, Snehasish; Rabaan, Ali A.; Al Mutair, Abbas; Alhumaid, Saad; Al-Tawfiq, Jaffar A.; Dhama, Kuldeep; Medicine, School of Medicine
    Objective: Internet of Things (IoT) integrates several technologies where devices learn from the experience of each other thereby reducing human-intervened likely errors. Modern technologies like IoT and machine learning enable the conventional to patient-specific approach transition in healthcare. In conventional approach, the biggest challenge faced by healthcare professionals is to predict a disease by observing the symptoms, monitoring the remote area patient, and also attending to the patient all the time after being hospitalised. IoT provides real-time data, makes decision-making smarter, and provides far superior analytics, and all these to help improve the quality of healthcare. The main objective of the work was to create an IoT-based automated system using machine learning models for symptom-based COVID-19 prognosis. Methods: Comparative analysis of predictive microbiology of COVID-19 from case symptoms using various machine learning classifiers like logistics regression, k-nearest neighbor, support vector machine, random forest, decision trees, Naïve Bayes, and gradient booster is reported here. For the sake of the validation and verification of the models, performance of each model based on the retrieved cloud-stored data was measured for accuracy. Results: From the accuracy plot, it was concluded that k-NN was more accurate (97.97%) followed by decision tree (97.79), support vector machine (97.42), logistics regression (96.50), random forest (90.66), gradient boosting classifier (87.77), and Naïve Bayes (73.50) in COVID-19 prognosis. Conclusion: The paper presents a health monitoring IoT framework having high clinical significance in real-time and remote healthcare monitoring. The findings reported here and the lessons learnt shall enable the healthcare system worldwide to counter not only this ongoing COVID but many other such global pandemics the humanity may suffer from time to come.