The analysis of longitudinal data using mixed model L-splines.

Sue J Welham; Brian R Cullis; Michael G Kenward; Robin Thompson

doi:10.1111/j.1541-0420.2005.00500.x

Home/Archive/PubMed

PubMed · 16918903

The analysis of longitudinal data using mixed model L-splines.

Sue J Welham·Brian R Cullis·Michael G Kenward·Robin Thompson

View original source ↗Export metadata ↓

Abstract

L-splines are a large family of smoothing splines defined in terms of a linear differential operator. This article develops L-splines within the context of linear mixed models and uses the resulting mixed model L-spline to analyze longitudinal data from a grassland experiment. In the spirit of time-series analysis, a periodic mixed model L-spline is developed, which partitions data into a smooth periodic component plus smooth long-term trend.

Explore related subjects

Biomass Biometry Data Interpretation, Statistical Likelihood Functions Linear Models Longitudinal Studies Models, Biological Models, Statistical Poaceae Journal Article Research Support, Non-U.S. Gov't

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Sue J Welham, Brian R Cullis, Michael G Kenward, Robin Thompson. 2006. The analysis of longitudinal data using mixed model L-splines.. https://doi.org/10.1111/j.1541-0420.2005.00500.x

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

Liquefaction of lignocellulose at high-solids concentrations.

To improve process economics of the lignocellulose to ethanol process a reactor system for enzymatic liquefaction and saccharification at high-solids concentrations was developed. The technology is based on free fall mixing employing a horizontally placed drum with a horizontal rotating shaft mounted with paddlers for mixing. Enzymatic liquefaction and saccharification of pretreated wheat straw was tested with up to 40% (w/w) initial DM. In less than 10 h, the structure of the material was changed from intact straw particles (length 1-5 cm) into a paste/liquid that could be pumped. Tests revealed no significant effect of mixing speed in the range 3.3-11.5 rpm on the glucose conversion after 24 h and ethanol yield after subsequent fermentation for 48 h. Low-power inputs for mixing are therefore possible. Liquefaction and saccharification for 96 h using an enzyme loading of 7 FPU/g.DM and 40% DM resulted in a glucose concentration of 86 g/kg. Experiments conducted at 2%-40% (w/w) initial DM revealed that cellulose and hemicellulose conversion decreased almost linearly with increasing DM. Performing the experiments as simultaneous saccharification and fermentation also revealed a decrease in ethanol yield at increasing initial DM. Saccharomyces cerevisiae was capable of fermenting hydrolysates up to 40% DM. The highest ethanol concentration, 48 g/kg, was obtained using 35% (w/w) DM. Liquefaction of biomass with this reactor system unlocks the possibility of 10% (w/w) ethanol in the fermentation broth in future lignocellulose to ethanol plants.

Biomass↗

PubMed2007

On maintenance models in severely and long-term limited membrane bioreactor cultivations.

Membrane bioreactors (MBRs) are combinations of common bioreactors and membrane separation units for biomass retention. Through increased biomass concentration, they allow increased productivity (or smaller reactor volume, respectively). Besides high biomass concentrations, operation at very low growth rates is typical for MBRs. In this regime, maintenance metabolism where substrate uptake only yields energy for cell survival becomes of higher importance than in processes run at higher growth rates. While thermodynamically based correlations for the prediction of maintenance coefficients are available for chemostat or other medium growth rate processes, some authors have mentioned a change in energy demand in MBRs and a dependence of maintenance parameters on operating conditions. Due to the fact that often mixed cultures are used and resulting from the different evaluation methods used by different authors, views on the possible influences on maintenance parameters differ. However, it is accepted that common models describing microbial growth and production of metabolites or degradation of pollutants do not consider the effects caused by severe limitations and therefore cannot sufficiently be applied to MBRs. In this study, maintenance parameters were determined for a model organism (Ustilago maydis) and results from different evaluation methods were compared. A continuous fit of respiration data gave more consistent results than the traditional method of plotting specific uptake versus growth rate. They suggest that below micro = 10% micro(max) the maintenance coefficient drops to a third of the value in short-term limited cultures.

Biomass↗

PubMed2007

Influence of pH, temperature, and urea molar flowrate on Arthrospira platensis fed-batch cultivation: a kinetic and thermodynamic approach.

Arthrospira platensis was cultivated photoautotrophically at 6.0 klux light intensity in 5.0-L open tanks, using a mineral medium containing urea as nitrogen source. Fed-batch experiments were performed at constant flowrate. A central composite factorial design combined to response surface methodology (RSM) was utilized to determine the relationship between the selected response variables (cell concentration after 10 days, X(m), cell productivity, P(X), and nitrogen-to-cell conversion factor, Y(X/N)) and codified values of the independent variables (pH, temperature, T, and urea flowrate, K). By applying the quadratic regression analysis, the equations describing the behaviors of these responses as simultaneous functions of the selected independent variables were determined, and the conditions for X(m) and P(X) optimization were estimated (pH 9.5, T = 29 degrees C, and K = 0.551 mM/day). The experimental data obtained under these conditions (X(m) = 749 mg/L; P(X) = 69.9 mg/L.day) were very close to the estimated ones (X(m) = 721 mg/L; P(X) = 67.1 mg/L.day). Additional cultivations were carried out under the above best conditions of pH control and urea flowrate at variable temperature. Consistently with the results of RSM, the best growth temperature was 29 degrees C. The maximum specific growth rates at different temperatures were used to estimate the thermodynamic parameters of growth (DeltaH* = 59.3 kJ/mol; DeltaS* = -0.147 kJ/mol.K; DeltaG* = 103 kJ/mol) and its thermal inactivation (DeltaH(D) (o) = 72.0 kJ/mol; DeltaS(D) (o) = 0.144 kJ/mol.K; DeltaG(D) (o) = 29.1 kJ/mol).

Biomass↗