GMMAT is an R package for performing genetic association tests in genome-wide association studies (GWAS) and sequencing association studies, for outcomes with distribution in the exponential family (e.g. binary outcomes) based on generalized linear mixed models (GLMMs). It can be used to analyze genetic data from individuals with population structure and relatedness. GMMAT fits a GLMM with covariate adjustment and random effects to account for population structure and familial or cryptic relatedness. For GWAS, GMMAT performs score tests for each genetic variant. For candidate gene studies, GMMAT can also perform Wald tests to get the effect size estimate for each genetic variant. For rare variant analysis from sequencing association studies, GMMAT performs the variant Set Mixed Model Association Tests (SMMAT), including the burden test, the sequence kernel association test (SKAT), SKAT-O and an efficient hybrid test of the burden test and SKAT, based on user-defined variant sets. See user manual here.References:
Evaluation of surveillance systems for early detection of outbreaks is particularly challenging when the systems are designed to detect events for which minimal or no historic examples exist (1). Although infection by biologic agents is rare, exceptions have occurred. For example, in 1979, persons living in Sverdlovsk in the former Soviet Union were exposed to Bacillus anthracis during an unintentional release from a weapons plant (2), and a limited number of persons were exposed in Florida, New York, and the District of Columbia during 2001 when B. anthracis spores were released through the mail (3). However, absent sufficient real outbreak data, measuring a system's detection performance requires simulation. Simulated outbreaks must reflect the diversity of threats, both natural and man-made, that a surveillance system might reasonably be expected to encounter and detect. This paper describes a flexible approach to generating standardized simulated data sets for benchmarking surveillance systems and provides examples of its application. Rather than model all possible conditions and factors, the approach relies on simulated outbreaks characterized by a controlled feature set that systematically defines the magnitude, temporal progression, duration, and spatial characteristics of the simulated outbreaks on the basis of variable parameters.Stages of Outbreak Detection
School Models Rare Sets
Performance of outbreak-detection models can be measured by using authentic data, synthetic data, or combinations of the two (Table). Two kinds of purely authentic data sets are possible. One is genuine syndromic data contemporaneous with either a known large-scale local outbreak (e.g., a winter influenza surge) (11) or a more circumscribed event (e.g., a diarrheal outbreak) (12). The data set would contain the background of ordinary disease or symptom occurrence and the signal of the actual outbreak. A second type of authentic data set is a hybrid containing background from a regional surveillance system spiked with cases from a known outbreak. This approach was taken when over-the-counter medication-sales data were spiked with an outbreak based on the Sverdlosk incident (13). Alternatively, a hypothetical baseline can be constructed, and actual or simulated signals can be imposed and injected. Although this approach is valid, limited need exists to simulate background activity, given the abundance of readily available real-signal streams from surveillance systems.
The approach described in this paper superimposes a simulated signal onto an authentic baseline, permitting exploration of the effects of controlled variations of signal characteristics. Two main approaches can be taken to creating this simulated signal: 1) using multistage, multivariate mathematical models to produce the signal or 2) defining a series of parameters that enable generation of a controlled feature set simulated signal. For example, a complex mathematical model (14) might be based on a scenario in which a particular form of aerosolized B. anthracis is dispersed under a certain set of atmospheric conditions over a specific geographic region with a well-characterized population demographic. The number of susceptible persons might be estimated and their subsequent behaviors modeled. The resulting effect on the syndromic surveillance data set (e.g., retail sales, primary care visits, or ED visits) could be projected. However, this approach for evaluating outbreak-detection performance is labor-intensive, and the models are based on multiple assumptions. A more flexible approach is to use a set of variable parameters describing a particular outbreak. Defining feature sets of outbreaks (e.g., magnitude, shape, and duration) allows rapid determination of the limits of a system's ability to detect an outbreak under varying conditions.Using Parameters To Specify Outbreak Characteristics
In the last set of experiments, the optimal method for integrating data from multiple regional EDs was determined (21). In one simulation, the synthetic outbreak was introduced evenly into both hospital data sets (aggregate model). In the second, the outbreak was introduced into only one or the other of the hospital data sets (local model). The aggregate model had a higher sensitivity for detecting outbreaks that were evenly distributed between the hospitals. However, for outbreaks that were localized to one facility, maintaining individual models for each location proved to be better. Given the complementary benefits offered by both approaches, the results suggested building a hybrid system that includes both individual models for each location and an aggregate model that combines all the data.Limitations
Lucille and Desi were well-known for their savvy as co-producers of I Love Lucy (they created their own roles for the show, which was rare), but Lucy began her career as a super-involved producer many years earlier. While studying drama in school, she got very involved in the creation of the plays.
Thanks for clarifying about the incidental parameters problem. I get your point about the criteria for MAR, that the missigness should not depend on the value of the datum. Key characteristics that could affect attrition are not observed in my data (e.g. SES, maternal characteristics, family income etc.). If there is no way to determine MAR, will it be fine to use a weighting procedure based on the theory of selection on observables ? For e.g. Fitzgerald and Moffit (1998) developed an indirect method to test attrition bias in panel data by using lagged outcomes to predict non-attrition. They call the lagged outcomes as auxillary variables. I ran probit regressions using different sets of lagged outcomes (such as lagged costs, hospitalization status, disability status etc.)and none of the models predicted >10% variation in non-attrition. This essentially means that attrition is probably not affected by observables. But should I still weight my observations in the panel regressions using the predicted probabilities of non-attrition from the probit models ?
I am fitting a discrete hazard model, so it feels strange not to specify clustered standard errors. In any case, firthlogit has produced results nearly identical to the results from logit and rare events logit models with clustered standard errors.
Dear Professor Allison,I have a hierarchical dataset consisting of three levels (N1=146,000; N2=402; N3=16). The dependent variable has 600 events. In my models I use a maximum of 12 predictors. I wonder whether your EPV rule of thumb also applies to a multilevel setting because up to now, following your rule, I apply a simple multilevel logistic regression.If not, are there any possibilities to correct for rare events in multilevel models (Firth-regression seems not to be available for multilevel logit, at least in Stata which I am using).
Lego outsourced production of the monorail tracks to another company, which went bankrupt and the tools to make the tracks were lost, so no more monorail sets could be made, which makes this set especially valuable. This set is exceedingly rare to find new in a box.
This is a list of BIONICLE sets considered the rarest. This is based on numerous factors, including their original status as limited edition sets, or their current availability/prominence on sites such as eBay and Bricklink. The list also assumes these sets are used, as unopened sets would be worth significantly more.
Based on their prices, 8935 Nocturn, 8940 Karzahni, 8953 Icarax and 8998 Toa Mata Nui are among the rarest of the sets listed above. Compared to their original retail prices, these sets are currently sold for three times their original worth. Other sets, like the limited edition 8942 Jetrax T6, are extremely hard to find in the first place.
Collections are ordered sets of models. You can bind "change" events to be notified when any model in the collection has been modified, listen for "add" and "remove" events, fetch the collection from the server, and use a full suite of Underscore.js methods.
Cristi Balan and Irina Dumitrascu created Tzigla, a collaborative drawing application where artists make tiles that connect to each other to create surreal drawings. Backbone models help organize the code, routers provide bookmarkable deep links, and the views are rendered with haml.js and Zepto. Tzigla is written in Ruby (Rails) on the backend, and CoffeeScript on the frontend, with Jammit prepackaging the static assets. 2ff7e9595c
Comments