Comparison of multiple comparison methods for identifying differential gene expression in simulated and real papillary thyroid cancer microarray data

The difficulty of detecting differential gene expression in microarray data has existed for many years. Several correction procedures try to avoid the family-wise error rate in multiple comparison process, including the Bonferroni and Sidak single-step p-value adjustments, Holm's step-down correction method, and Benjamini and Hochberg's false discovery rate (FDR) correction procedure. Each multiple comparison technique has its advantages and weaknesses. We studied each multiple comparison method through numerical studies (simulations) and applied the methods to the real exploratory DNA microarray data, which detect of molecular signatures in papillary thyroid cancer (PTC) patients. According to our results of simulation studies, Benjamini and Hochberg step-up FDR controlling procedure is the best process among these multiple comparison methods and we discovered 1277 potential biomarkers among 54675 probe sets after applying the Benjamini and Hochberg's method to PTC microarray data.^

Extension of k-ratio methods to rank-based analyses

An extension of k-ratio multiple comparison methods to rank-based analyses is described. The new method is analogous to the Duncan-Godbold approximate k-ratio procedure for unequal sample sizes or correlated means. The close parallel of the new methods to the Duncan-Godbold approach is shown by demonstrating that they are based upon different parameterizations as starting points.^ A semi-parametric basis for the new methods is shown by starting from the Cox proportional hazards model, using Wald statistics. From there the log-rank and Gehan-Breslow-Wilcoxon methods may be seen as score statistic based methods.^ Simulations and analysis of a published data set are used to show the performance of the new methods. ^

A COMPARISON OF TWO METHODS FOR THE DETECTION OF FECAL POLLUTION IN DRINKING WATER (HYDROGEN SULFIDE, INDICATOR, MODIFICATION)

A new technique for the detection of microbiological fecal pollution in drinking and in raw surface water has been modified and tested against the standard multiple-tube fermentation technique (most-probable-number, MPN). The performance of the new test in detecting fecal pollution in drinking water has been tested at different incubation temperatures. The basis for the new test was the detection of hydrogen sulfide produced by the hydrogen sulfide producing bacteria which are usually associated with the coliform group. The positive results are indicated by the appearance of a brown to black color in the contents of the fermentation tube within 18 to 24 hours of incubation at 35 (+OR-) .5(DEGREES)C. For this study 158 water samples of different sources have been used. The results were analyzed statistically with the paired t-test and the one-way analysis of variance. No statistically significant difference was noticed between the two methods, when tested 35 (+OR-) .5(DEGREES)C, in detecting fecal pollution in drinking water. The new test showed more positive results with raw surface water, which could be due to the presence of hydrogen sulfide producing bacteria of non-fecal origin like Desulfovibrio and Desulfomaculum. The survival of the hydrogen sulfide producing bacteria and the coliforms was also tested over a 7-day period, and the results showed no significant difference. The two methods showed no significant difference when used to detect fecal pollution at a very low coliform density. The results showed that the new test is mostly effective, in detecting fecal pollution in drinking water, when used at 35 (+OR-) .5(DEGREES)C. The new test is effective, simple, and less expensive when used to detect fecal pollution in drinking water and raw surface water at 35 (+OR-) .5(DEGREES)C. The method can be used for qualitative and/or quantitative analysis of water in the field and in the laboratory. ^

Comparison of alternative methods of contact tracing in a syphilis control program

The purpose of this study was to compare the relative effectiveness of alternative methods of tracing named contacts of syphilis patients. A total of 236 contacts, identified by patients in two City of Houston Department of Health and Human Services clinics during the period April 1 through July 31, 1987, were studied. After contacts were grouped by sex and age, the proportion brought to examination by each of three methods, and by a combination of methods, was determined for each subgroup.^ The study found that 78.4% of all the 236 named sex contacts reported were located and brought to examination by the various methods of contact tracing and that 21.6% were missed. Of the 185 contacts examined, a combination of methods identified 47.7% of the cases, telephone contact, 28.6%, field contact, 16.9%, and patient referral, 11.8%.^ Of the 236 contacts reported, males made up 56.8% and females 43.2%. Contact tracing was more successful among females, with 81.4% of the 102 named female contacts, as compared to 76.1% of the 134 named male contacts being brought to examination. It is not known whether equal efforts were exerted in the follow-up of both male and female contacts. In both female and male subgroups, a combination of methods brought over 40% of sex contacts to examination. Telephone contact among females yielded 27.7% of the cases and field contact 18.1%, whereas in males, telephone contact identified 29.4% of the cases and field contact 15.7%. Patient referral was the least productive method in both sex groups, locating 12.8% in males as compared to 10.8% in females.^ On an age specific basis, a combination of methods was the most effective method in the 15-39 age group, whereas telephone contact was most effective in the 40-44 age group, and field contact in the 50-54 age group. Of all the methods of contact tracing, patient referral was the least productive in most age groups.^ Future studies of contact tracing should incorporate several important variables which were not examined in this study. ^

Comparison of methods for assessing in home night time crying in healthy infants

Background. Infant colic is a common condition that is thought to put infants at risk for Shaken Baby Syndrome, a particularly devastating form of child abuse. However, little research has been done on techniques parents can use to deal with infant colic. This pilot study was conducted to assess the equipment that will be used in a randomized control trial that will compare the results for two different techniques that parents can use to reduce crying in infants with colic. ^ Methods. A total of 11 healthy infants, between one and five months of age, were recruited into this pilot study. All infants had a dosimeter, actiwatch and maternal log placed into the home and a subset of infants (N=3) were also recorded by a video camera. The equipment recorded between 6pm and 6am for at least two and up to five nights. The maternal log and video log were compared with one another to determine if the maternal log provides an accurate representation of the infant's night-time activities (i.e. sleep, awake, crying, feeding). The maternal log was then compared to the dosimeter and actiwatch data to determine if the dosimeter/actiwatch accurately reproduce the maternal log. ^ Results. Data from 10 infants were included in the analyses. The maternal log and video log were in full or partial agreement 90% of the time. When comparing events noted by the mother, the maternal log and dosimeter data were in agreement 84% of the time, and the maternal log and actiwatch data were in agreement 87% of the time. In combination, the dosimeter and/or actiwatch data agreed with the maternal log 90% of the time. ^ Conclusions. Our preliminary analyses of these data suggest the dosimeter and actiwatch will be useful tool for defining infant sleep patterns relative to the maternal log. However further analysis will be required to develop threshold values that can be used to objectively define events in the proposed RCT. Such analyses will need to integrate data from multiple dosimeters and deal with the shifting baselines observed for both the dosimeter and actiwatch.^

A comparison of two statistical methods for analysis of day of the week effect on births from 2005-2007 in Harris County of Texas

The pattern of the births during the week has been reported by many studies. The births occurred in weekends are found consistently less then births occurred in weekdays. This study employed two statistical methods, two-way ANOVA and two-way Friedman's test to analyse the daily variations in amount of births of 222,735 births from 2005-2007 in Harris County, Texas. The two methods were compared on their assumptions, procedures and results. Both of the tests showed a significant result which indicated that the births through the week are not uniformly distributed. The result of multiple comparison demonstrated the births occurring on weekends were significantly different than the births occurring on weekdays with least amount on Sundays.^

Comparison of two statistical methods for rare variant association analysis

Studies have shown that rare genetic variants have stronger effects in predisposing common diseases, and several statistical methods have been developed for association studies involving rare variants. In order to better understand how these statistical methods perform, we seek to compare two recently developed rare variant statistical methods (VT and C-alpha) on 10,000 simulated re-sequencing data sets with disease status and the corresponding 10,000 simulated null data sets. The SLC1A1 gene has been suggested to be associated with diastolic blood pressure (DBP) in previous studies. In the current study, we applied VT and C-alpha methods to the empirical re-sequencing data for the SLC1A1 gene from 300 whites and 200 blacks. We found that VT method obtains higher power and performs better than C-alpha method with the simulated data we used. The type I errors were well-controlled for both methods. In addition, both VT and C-alpha methods suggested no statistical evidence for the association between the SLC1A1 gene and DBP. Overall, our findings provided an important comparison of the two statistical methods for future reference and provided preliminary and pioneer findings on the association between the SLC1A1 gene and blood pressure.^

Comparison of machine learning approaches and statistical methods for estimation of airborne chemical exposures

Accurate quantitative estimation of exposure using retrospective data has been one of the most challenging tasks in the exposure assessment field. To improve these estimates, some models have been developed using published exposure databases with their corresponding exposure determinants. These models are designed to be applied to reported exposure determinants obtained from study subjects or exposure levels assigned by an industrial hygienist, so quantitative exposure estimates can be obtained. ^ In an effort to improve the prediction accuracy and generalizability of these models, and taking into account that the limitations encountered in previous studies might be due to limitations in the applicability of traditional statistical methods and concepts, the use of computer science- derived data analysis methods, predominantly machine learning approaches, were proposed and explored in this study. ^ The goal of this study was to develop a set of models using decision trees/ensemble and neural networks methods to predict occupational outcomes based on literature-derived databases, and compare, using cross-validation and data splitting techniques, the resulting prediction capacity to that of traditional regression models. Two cases were addressed: the categorical case, where the exposure level was measured as an exposure rating following the American Industrial Hygiene Association guidelines and the continuous case, where the result of the exposure is expressed as a concentration value. Previously developed literature-based exposure databases for 1,1,1 trichloroethane, methylene dichloride and, trichloroethylene were used. ^ When compared to regression estimations, results showed better accuracy of decision trees/ensemble techniques for the categorical case while neural networks were better for estimation of continuous exposure values. Overrepresentation of classes and overfitting were the main causes for poor neural network performance and accuracy. Estimations based on literature-based databases using machine learning techniques might provide an advantage when they are applied to other methodologies that combine `expert inputs' with current exposure measurements, like the Bayesian Decision Analysis tool. The use of machine learning techniques to more accurately estimate exposures from literature-based exposure databases might represent the starting point for the independence from the expert judgment.^