Assessment of the effect on statistical power of regression model misspecification by using techniques of mathematical statistics and simulation study

Objectives. This paper seeks to assess the effect on statistical power of regression model misspecification in a variety of situations. ^ Methods and results. The effect of misspecification in regression can be approximated by evaluating the correlation between the correct specification and the misspecification of the outcome variable (Harris 2010).In this paper, three misspecified models (linear, categorical and fractional polynomial) were considered. In the first section, the mathematical method of calculating the correlation between correct and misspecified models with simple mathematical forms was derived and demonstrated. In the second section, data from the National Health and Nutrition Examination Survey (NHANES 2007-2008) were used to examine such correlations. Our study shows that comparing to linear or categorical models, the fractional polynomial models, with the higher correlations, provided a better approximation of the true relationship, which was illustrated by LOESS regression. In the third section, we present the results of simulation studies that demonstrate overall misspecification in regression can produce marked decreases in power with small sample sizes. However, the categorical model had greatest power, ranging from 0.877 to 0.936 depending on sample size and outcome variable used. The power of fractional polynomial model was close to that of linear model, which ranged from 0.69 to 0.83, and appeared to be affected by the increased degrees of freedom of this model.^ Conclusion. Correlations between alternative model specifications can be used to provide a good approximation of the effect on statistical power of misspecification when the sample size is large. When model specifications have known simple mathematical forms, such correlations can be calculated mathematically. Actual public health data from NHANES 2007-2008 were used as examples to demonstrate the situations with unknown or complex correct model specification. Simulation of power for misspecified models confirmed the results based on correlation methods but also illustrated the effect of model degrees of freedom on power.^

Protein quaternary structure determination using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and chemical crosslinking

A means of analyzing protein quaternary structure using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI MS) and chemical crosslinking was evaluated. Proteins of known oligomeric structure, as well as monomeric proteins, were analyzed to evaluate the method. The quaternary structure of proteins of unknown or uncertain structure was investigated using this technique. The stoichiometry of recombinant E. coli carbamoyl phosphate synthetase and recombinant human farnesyl protein transferase were determined to be heterodimers using glutaraldehyde crosslinking, agreeing with the stoichiometry found for the wild type proteins. The stoichiometry of the gamma subunit of E. coli DNA polymerase III holoenzyme was determined in solution without the presence of other subunits to be a homotetramer using glutaraldehyde crosslinking and MALDI MS analysis. Chi and psi subunits of E. coli DNA polymerase III subunits appeared to form a heterodimer when crosslinked with heterobifunctional photoreactive crosslinkers.^ Comparison of relative % peak areas obtained from MALDI MS analysis of crosslinked proteins and densitometric scanning of silver stained sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) gels showed excellent qualitative agreement for the two techniques, but the quantitative analyses differed, sometimes significantly. This difference in quantitation could be due to SDS-PAGE conditions (differential staining, loss of sample) or to MALDI MS conditions (differences in ionization and/or detection). Investigation of pre-purified crosslinked monomers and dimers recombined in a specific ratio revealed the presence of mass discrimination in the MALDI MS process. The calculation of mass discrimination for two different MALDI time-of-flight instruments showed the loss of a factor of approximately 2.6 in relative peak area as the m/z value doubles over the m/z range from 30,000 to 145,000 daltons.^ Indirect symmetry was determined for tetramers using glutaraldehyde crosslinking with MALDI MS analysis. Mathematical modelling and simple graphing allowed the determination of the symmetry for several tetramers known to possess isologous D2 symmetry. These methods also distinguished tetramers that did not fit D2 symmetry such as apo-avidin. The gamma tetramer of E. coli DNA polymerase III appears to have isologous D2 symmetry. ^

Systematic review of genetic risk score in coronary heart disease and other diseases

In order to better take advantage of the abundant results from large-scale genomic association studies, investigators are turning to a genetic risk score (GRS) method in order to combine the information from common modest-effect risk alleles into an efficient risk assessment statistic. The statistical properties of these GRSs are poorly understood. As a first step toward a better understanding of GRSs, a systematic analysis of recent investigations using a GRS was undertaken. GRS studies were searched in the areas of coronary heart disease (CHD), cancer, and other common diseases using bibliographic databases and by hand-searching reference lists and journals. Twenty-one independent case-control studies, cohort studies, and simulation studies (12 in CHD, 9 in other diseases) were identified. The underlying statistical assumptions of the GRS using the experience of the Framingham risk score were investigated. Improvements in the construction of a GRS guided by the concept of composite indicators are discussed. The GRS will be a promising risk assessment tool to improve prediction and diagnosis of common diseases.^