With technological advances now allowing measurement of thousands of genes, proteins and metabolites, researchers are using this information to develop diagnostic and prognostic tests and discern the biological pathways underlying diseases. Often, an investigator's objective is to develop a classification rule to predict group membership of unknown samples based on a small set of features and that could ultimately be used in a clinical setting. While common classification methods such as random forest and support vector machines are effective at separating groups, they do not directly translate into a clinically-applicable classification rule based on a small number of features.We present a simple feature selection and classification method for biomarker detection that is intuitively understandable and can be directly extended for application to a clinical setting. We first use a jackknife procedure to identify important features and then, for classification, we use voting classifiers which are simple and easy to implement. We compared our method to random forest and support vector machines using three benchmark cancer 'omics datasets with different characteristics. We found our jackknife procedure and voting classifier to perform comparably to these two methods in terms of accuracy. Further, the jackknife procedure yielded stable feature sets. Voting classifiers in combination with a robust feature selection method such as our jackknife procedure offer an effective, simple and intuitive approach to feature selection and classification with a clear extension to clinical applications.

Background

Severe burn and traumatic brain injuries (TBI) lead to significant mortality, and combined burn-TBI injuries may predispose towards even worse outcomes. The purpose of this study was to investigate the mortality of patients with burn, burn with non-TBI trauma, and combined burn/TBI to determine if combined injury portends a worse outcome.

Methods

We obtained the National Trauma Data Bank from 2007 to 2012, identifying 32,334 patients with burn related injuries, dividing this cohort into three injury types: BURN ONLY, BURN with TRAUMA/NO TBI, and BURN with TBI. For each patient, demographic data was obtained, including age, gender, presence of trauma, TBI, or inhalation injury, burn total body surface area (TBSA), Glasgow Coma Scale, Injury Severity Score, and mortality. Multivariable logistic regression was performed.

Results

Age, gender, and TBSA were similar across the three injury groups, but the incidence of inhalation injury was doubled in the BURN with TRAUMA/NO TBI (15.4 %) and BURN with TBI (15.3 %) groups when compared to the BURN ONLY (7.2 %) group. Mortality differed across injury categories after adjusting for age, TBSA, and inhalation injury. Increased mortality was seen in BURN with TRAUMA/NO TBI versus BURN ONLY (OR = 1.27 [1.06, 1.53]) and was higher when comparing BURN with TBI versus BURN ONLY (OR = 4.22 [2.85, 6.18]). BURN with TBI also had higher mortality when compared to BURN with TRAUMA/NO TBI (OR = 3.33 [2.30, 4.82]). The logs odds of mortality also increased with increasing age, TBSA and presence of inhalation injury.

Discussion

This analysis of the NTDB suggests that mortality following burn-related injuries may be higher when burn injury is combined with TBI when compared to burns with other trauma, even after correcting for age, TBSA, and inhalation injury. Further clinical and laboratory research is needed to validate these findings and better understand how to optimize combined TBI and burn injury treatment.

American Indian/Alaska Native (AI/AN) women have the lowest cancer-screening rate of any ethnic or racial group; AI/AN women in all regions are less likely than non-Hispanic white women to be diagnosed with localized breast cancer; and those AI/AN women presenting with breast cancer have the lowest 5-year survival rate compared to other ethnic groups. This study found that cultural beliefs are more of a factor in mammography screening behavior than other barriers such as access; and that a more holistic educational intervention designed by AI/AN women prompted individual intent and actions to seek mammograms among AI/AN women >40 and to change unhealthy eating and sedentary lifestyles.

Missing values are common in high-throughput mass spectrometry data. Two strategies are available to address missing values: (i) eliminate or impute the missing values and apply statistical methods that require complete data and (ii) use statistical methods that specifically account for missing values without imputation (imputation-free methods). This study reviews the effect of sample size and percentage of missing values on statistical inference for multiple methods under these two strategies. With increasing missingness, the ability of imputation and imputation-free methods to identify differentially and non-differentially regulated compounds in a two-group comparison study declined. Random forest and k-nearest neighbor imputation combined with a Wilcoxon test performed well in statistical testing for up to 50% missingness with little bias in estimating the effect size. Quantile regression imputation accompanied with a Wilcoxon test also had good statistical testing outcomes but substantially distorted the difference in means between groups. None of the imputation-free methods performed consistently better for statistical testing than imputation methods.

Background

Effective management depends upon accurately estimating trends in abundance of bird populations over time, and in some cases estimating abundance. Two population estimation methods, double observer (DO) and double sampling (DS), have been advocated for avian population studies and the relative merits and short-comings of these methods remain an area of debate.

Methodology/principal findings

We used simulations to evaluate the performances of these two population estimation methods under a range of realistic scenarios. For three hypothetical populations with different levels of clustering, we generated DO and DS population size estimates for a range of detection probabilities and survey proportions. Population estimates for both methods were centered on the true population size for all levels of population clustering and survey proportions when detection probabilities were greater than 20%. The DO method underestimated the population at detection probabilities less than 30% whereas the DS method remained essentially unbiased. The coverage probability of 95% confidence intervals for population estimates was slightly less than the nominal level for the DS method but was substantially below the nominal level for the DO method at high detection probabilities. Differences in observer detection probabilities did not affect the accuracy and precision of population estimates of the DO method. Population estimates for the DS method remained unbiased as the proportion of units intensively surveyed changed, but the variance of the estimates decreased with increasing proportion intensively surveyed.

Conclusions/significance

The DO and DS methods can be applied in many different settings and our evaluations provide important information on the performance of these two methods that can assist researchers in selecting the method most appropriate for their particular needs.

Background/Study Context: A number of longitudinal randomized controlled trials (LRCT) have used free verbal recall tests to study the effects of post-menopausal estrogen hormone therapy (HT) on episodic memory, but none have explicitly explored contrasts between list and story recall, in spite of cognitive differences between the tasks. For example, list recall provides little support for the use of gist, while story recall emphasizes it, and there is evidence that estrogen produces gist bias. Moreover, we present a literature tabulation that also suggests a task-specific HT effect.

Methods

In an LRCT with up to eight yearly test sessions, post-menopausal women were randomly assigned either to placebo (N = 56) or to an estrogen formulation (N = 44); subgroups received either estrogen alone (hysterectomy; E-alone; N = 16) or with progestin (intact uterus; E + P; N = 28). Participants were tested on the immediate and delayed list and story recall at each session.

Results

Linear mixed effects analyses of longitudinal trajectories showed that relative to placebo, the HT group declined significantly faster on immediate list recall and slower on immediate story recall. Separate analyses produced a sharpened version of this pattern for the E-alone subgroup but found no significant effects for the E + P subgroup. No significant effects were found in delayed testing.

Conclusion

The dissociation we found for immediate list and story recall is similar to the pattern of results in our literature tabulation. Fuzzy-Trace Theory posits parallel verbatim and gist traces plus a meta-cognitive review which becomes more gist-biased with age. Our results suggest that: (1) estrogen increases gist bias, hastening the normal age-related decline of list recall but slowing the decline of story recall relative to placebo; (2) decay of the verbatim trace over time generally causes a shift to gist, thereby accounting for the absence of a delayed recall difference; and (3) progestin weakens the effects of estrogen, thereby accounting for why the dissociation found in E-alone was absent in the E + P subgroup.

BACKGROUND: Sliding hiatal herniation (SHH) and gastroesophageal reflux (GER) commonly occur in French bulldogs. Wireless pH monitoring can quantitatively assess acidic GER in dogs affected by SHH. HYPOTHESIS/OBJECTIVES: Measure acidic GER in French bulldogs with SHH, pre- and post-brachycephalic obstructive airway syndrome (BOAS) surgery, utilizing a wireless pH capsule (Bravo Calibration-free, Medtronic, Minnesota), and correlate with owners observations of regurgitation. ANIMALS: Eleven French bulldogs diagnosed with SHH via swallowing fluoroscopy. METHODS: Prospective cohort study. A pH capsule was endoscopically placed in the esophagus. Up to 96 hours of data were acquired as the owner logged clinical signs. Spearmans correlation and Wilcoxon rank-sum tests evaluated factors correlated with acid exposure time (AET), defined by the % time pH < 4. In 4/11 dogs, Bravo monitoring was repeated 2-4 months after BOAS surgery. RESULTS: Medians (Q1-Q3) for age and weight were 21 months (17-35.5) and 10.0 kg (8.9-11.5). BOAS severity was mild (3), moderate (4), or severe (4). Medians (Q1-Q3) for AET and reflux events were 3.3% (2.6-6.4) and 70 (34-173). Clinical score (P = .82) and BOAS severity (P = .60) were not correlated with AET, but age was negatively correlated (rho = -.66, P = .03). Median probability (Q1-Q3) that regurgitation was associated with a reflux event was 72.5% (0-99). Percent AET numerically improved in all 4 dogs that underwent BOAS surgery although not statistically assessed. CONCLUSIONS AND CLINICAL IMPORTANCE: Wireless pH monitoring documented acidic GER in French bulldogs with SHH, captured subclinical events, and showed improvements after BOAS surgery.

Objective

To investigate the effects of 3 different bowel preparation protocols on interpretation of computed tomography (CT) pneumocolonography images.

Study design

Experimental crossover design.

Animals

Intact male, hound-cross, research colony dogs (n=4).

Methods

All dogs received the 3 different bowel preparation protocols for CT pneumocolonography in the same order, with a minimum of 2 weeks between protocols. For each segment of large bowel, the subjective adequacy of bowel cleansing was assessed, residual fecal and bowel volumes were calculated, and the density of fecal material in the bowel lumen was measured. Linear mixed effect models that included a random dog effect were used to evaluate mean differences in outcome measures across protocols.

Results

No dogs experienced any clinical problems associated with the protocols or CT pneumocolonography. Bowel cleansing was considered adequate for CT pneumocolonography interpretation for all 3 protocols. There was a significant effect of protocol on residual fecal volumes and the fecal:bowel volume ratio, with the 2 protocols that included an extended fast producing the lowest total residual fecal volumes. There was a significant effect of protocol on maximum measured density of residual fecal material with the 2 protocols including iodinated contrast having the highest density.

Conclusions

All protocols were sufficient for CT pneumocolonography interpretation and contrast-tagging of residual fecal material was successful with oral iopamidol administration. An at-home bowel cleansing protocol may provide adequate bowel cleansing for CT pneumocolonography image interpretation.

Mass spectrometry is an important high-throughput technique for profiling small molecular compounds in biological samples and is widely used to identify potential diagnostic and prognostic compounds associated with disease. Commonly, this data generated by mass spectrometry has many missing values resulting when a compound is absent from a sample or is present but at a concentration below the detection limit. Several strategies are available for statistically analyzing data with missing values. The accelerated failure time (AFT) model assumes all missing values result from censoring below a detection limit. Under a mixture model, missing values can result from a combination of censoring and the absence of a compound. We compare power and estimation of a mixture model to an AFT model. Based on simulated data, we found the AFT model to have greater power to detect differences in means and point mass proportions between groups. However, the AFT model yielded biased estimates with the bias increasing as the proportion of observations in the point mass increased while estimates were unbiased with the mixture model except if all missing observations came from censoring. These findings suggest using the AFT model for hypothesis testing and mixture model for estimation. We demonstrated this approach through application to glycomics data of serum samples from women with ovarian cancer and matched controls.