PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Linear Regression”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 343 records · Page 19Linked to original sources
PubMed1998

On the use of linear regression and maximum likelihood for QTL mapping in half-sib designs.

Methods of identification of quantitative trait loci (QTL) using a half-sib design are generally based on least-squares or maximum likelihood approaches. These methods differ in the genetical model considered and in the information used. Despite these differences, the power of the two methods in a daughter design in very similar. Using an analogy with a one-way analysis of variance, we propose an equation connecting the two test-statistics (F ratio for regression and likelihood ratio test in the case of the maximum likelihood). The robustness of this relationship is tested by simulation for different single QTL models. In general, the correspondence between the two statistics is good under both the null hypothesis and the alternative hypothesis of a single QTL segregating. Practical implications are discussed with particular emphasis on the theoretical distribution of the likelihood ratio test.

Chromosome Mapping↗
PubMed1976

The significance for breeding of linear regression analysis of genotype-environment interactions.

Methods of regression analysis of genotype-environment interaction are considered in relation to existing theory dealing with the relative efficiencies of selection for general or specific adaptation to the environment, and the choice of environments for assessment. The two alternative models is involving regression on to environmental effects (model 2) or genotypic effects (model 3) are equivalent when regression lines are concurrent, but are shown to be mutually exclusive when concurrence is absent...

Crosses, Genetic↗
PubMed2001

Robust linear regression taking into account errors in the predictor and response variables.

We developed a robust regression technique that is a generalization of the least median of squares (LMS) technique to the field in which the errors in both the predictor and the response variables are taken into account. This simple generalization is limited in the sense that the resulting straight line is found by using only two points from the initial data set. In this way a simulation step is added by using the Monte Carlo method to generate the best robust regression line. We call this new technique 'bivariate least median of squares' (BLMS), following the notation of the LMS method. We checked the robustness of the new regression technique by calculating its breakdown point, which was 50%. This confirms the robustness of the BLMS regression line. In order to show its applicability to the chemical field we tested it on simulated data sets and real data sets with outliers. The BLMS robust regression line was not affected by many types of outlying points in the data sets.

Journal Article↗
PubMed1986

Quantitative electron microscope autoradiography: application of multiple linear regression analysis.

A new method for the analysis of high resolution EM autoradiographs is described. It identifies labelled cell organelle profiles in sections on a strictly statistical basis and provides accurate estimates for their radioactivity without the need to make any assumptions about their size, shape and spatial arrangement. The radioactivity in interfacial membranes and transitional junctional regions between pairs of adjoining cell structures can be determined without any additional measurements. The uniformity of internal labelling of large cell organelles can be also assessed. Correcting for cross-fire does not need any predictive information from a frequency distribution function describing the image spread about a radioactive point source. Instead, the section area is subdivided into regions in such a way that each region with cross-fired silver grains over it includes also the sites of the radioactive disintegrations that have produced them. The method is based on the quantitative regression relationship between the number of developed silver grains overlying any given region and the size of the cell organelle profiles included. A transparent overlay screen bearing a regular array of circles and a point-counting procedure are applied to divide the section area into regions and to measure their area and the size of the cell organelle profiles. The least squares method is used for the simple and rapid calculation of specific activity estimates and exact standard errors to be attached to them. A modest computing facility and a standard library program are required. Simulation modelling and analysis of residuals were used to lend support to the validity of the method.

Animals↗
PubMed1975

Conformational analysis of circular dichroism spectra of insulin, proinsulin and c-peptides by non-linear regression.

A method of resolving CD spectra in alpha-helix, beta-structure and random coil conformations is described. The residue ellipticites for alpha-helix and beta-structure given by Greenfield & Fasman or by Chen,, Yang & Martinez are used together with CD spectra from at least two similar peptides to determine, by an iterative least-squares method, the number of amino acids in the three reference conformations as well as a set of residue ellipticities characteristic of the random coils of the family of peptides in question, but not necessarily of other peptides. The fits between computed and experimental spectra improve significantly and systematic deviations disappear by allowing the random coil coefficients to vary from one family of proteins to another, a liberty justified by the different types of random coils that have been encountered. The method of analysis showed that 5 M urea did not change the conformations of C-peptides of proinsulin from ox, pig and duck, all being mainly in the random coil conformation and all having 3-4 amino acids in beta-structure. Bovine insulin and proinsulin showed a transfer of amino acids from alpha-helix to beta-structure with increasing concentrations of urea, the latter at a higher concentration, indicating a stabilizing effect of the connecting peptide. The numbers of amino acids found in the alpha-helical conformation in insulin and proinsulin were equal and in agreement with the X-ray crystallographic data for insulin when the Greenfield & Fasman coefficients for alpha-helix and beta-structure were employed, whereas the Chen, Yang & Martinez coefficients yielded too few amino acids in alpha-helix in proinsulin. Both sets of coefficients estimate more beta-structure in proinsulin than in insulin.

Amino Acids↗