A conditional density error model for the statistical analysis of microarray data.

Brad Love; David R Rank; Sharron G Penn; David A Jenkins; Russell S Thomas

doi:10.1093/bioinformatics/18.8.1064

Home/Archive/PubMed

PubMed · 12176829

A conditional density error model for the statistical analysis of microarray data.

Brad Love·David R Rank·Sharron G Penn·David A Jenkins·Russell S Thomas

View original source ↗Export metadata ↓

Abstract

MOTIVATION: In many microarray experiments, relatively few intra- and inter-array replicate measurements are made due to significant cost limitations and sample availability. Compounding this problem is a lack of robust statistical methods for analyzing gene expression data with limited experimental replicates. As a result, the interpretation of the results of these experiments are difficult with little understanding of the probability of type I and type II errors. RESULTS: The variability in a series of replicate microarray measurements was modelled using a combination of parametric and non-parametric methods. A 3-dimensional surface was created for the conditional distribution of the variability given the mean signal intensity in both the Cy3 and Cy5 channels. The results were used as the basis for developing statistical methods for analyzing gene expression data with limited experimental replicates. AVAILABILITY: The statistical analysis scripts are available upon request.

Explore related subjects

Computer Simulation DNA DNA Replication Gene Expression Gene Expression Profiling Gene Expression Regulation Humans Models, Genetic Models, Statistical Oligonucleotide Array Sequence Analysis Reproducibility of Results Sensitivity and Specificity Comparative Study Journal Article

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

Brad Love, David R Rank, Sharron G Penn, David A Jenkins, Russell S Thomas. 2002. A conditional density error model for the statistical analysis of microarray data.. https://doi.org/10.1093/bioinformatics%2F18.8.1064

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

KEEP EXPLORING

Related citations

PubMed2025

Generating correlated data for omics simulation.

Simulation of realistic omics data is a key input for benchmarking studies that help users obtain optimal computational pipelines. Omics data involves large numbers of measured features on each sample and these measures are generally correlated with each other. However, simulation too often ignores these correlations, perhaps due to computational and statistical hurdles of doing so. To alleviate this, we describe three approaches for generating omics-scale data with correlated measures which mimic real datasets. These approaches are all based on a Gaussian copula approach with a covariance matrix that decomposes into a diagonal part and a low-rank part. This decomposition allows for extremely efficient simulation, overcoming a hurdle for adoption of past methods. We use these approaches to demonstrate the importance of including correlation in two benchmarking applications. First, we show that variance of results from the popular DESeq2 method increases when dependence is included. Second, we demonstrate that CYCLOPS, a method for inferring circadian time of collection from transcriptomics, improves in performance when given gene-gene dependencies in some circumstances. We provide an R package, dependentsimr, that has efficient implementations of these methods and can generate dependent data with arbitrary marginal distributions, including discrete (binary, ordered categorical, Poisson, negative binomial), continuous (normal), or with an empirical distribution.

Computer Simulation↗

PubMed2017

Addressing current challenges in cancer immunotherapy with mathematical and computational modelling.

The goal of cancer immunotherapy is to boost a patient's immune response to a tumour. Yet, the design of an effective immunotherapy is complicated by various factors, including a potentially immunosuppressive tumour microenvironment, immune-modulating effects of conventional treatments and therapy-related toxicities. These complexities can be incorporated into mathematical and computational models of cancer immunotherapy that can then be used to aid in rational therapy design. In this review, we survey modelling approaches under the umbrella of the major challenges facing immunotherapy development, which encompass tumour classification, optimal treatment scheduling and combination therapy design. Although overlapping, each challenge has presented unique opportunities for modellers to make contributions using analytical and numerical analysis of model outcomes, as well as optimization algorithms. We discuss several examples of models that have grown in complexity as more biological information has become available, showcasing how model development is a dynamic process interlinked with the rapid advances in tumour-immune biology. We conclude the review with recommendations for modellers both with respect to methodology and biological direction that might help keep modellers at the forefront of cancer immunotherapy development.

Computer Simulation↗

PubMed2003

Modeling alpha-helical coiled-coil interactions: the axial and azimuthal alignment of 1B segments from vimentin intermediate filaments.

Attempts at predicting the relative axial alignments of fibrous protein molecules in filamentous structures have relied upon representing the (multichain) molecular structure by a one-dimensional sequence of amino acids. Potential intermolecular ionic and apolar interactions were counted and determined as a function of the relative axial stagger between the molecules. No attempts were made to consider the azimuthal aspect of the interacting molecules and neither were apolar or ionic energy terms used. Surprisingly, this simple approach proved remarkably informative and yielded accurate predictions of the axial periods present. However, a more comprehensive analysis involving the energetics of aggregation taking due regard for the relative azimuths of the molecules as well as their separation should decrease the noise level in the calculations and reveal other pertinent information. Toward that end, we have modeled the interaction between two alpha-helical coiled-coil segments in intermediate filament molecules (1B segments from human vimentin). The relative axial alignment and polarity of the molecules is already known from detailed crosslinking studies and this provides a criterion against which the success (or otherwise) of the modeling can be judged. The results confirm that an antiparallel alignment of two 1B segments is preferred over any of the parallel options (as observed experimentally). The calculated axial alignment, however, is not identical to that observed from detailed crosslinking studies indicating that other parts of the molecule (probably the head and tail domains as well as other coiled-coil segments) have a crucial role in determining the precise mode of axial aggregation. The results also show that the apolar interactions seem to be significantly less important in the alignment process than the ionic ones. This is consistent with the observation of a well-defined period in the linear disposition of the charged (but not apolar) residues along the length of the outer surface of the vimentin molecule.

Computer Simulation↗