Susan R. Wilson

Two guidelines for bootstrap hypothesis testing

Two guidelines for nonparametric bootstrap hypothesis testing are highlighted. The first recommends that resampling be done in a way that reflects the null hypothesis, even when the true hypothesis is distant from the null. The second guideline argues that bootstrap hypothesis tests should employ methods that are already recognized as having good features in the closely related problem of confidence interval construction. Violation of the first guideline can seriously reduce the power of a test. Sometimes this reduction is spectacular, since it is most serious when the null hypothesis is grossly in error. The second guideline is of some importance when the conclusion of a test is equivocal. It has no direct bearing on power, but improves the level accuracy of a test.

Efficient experimental design and analysis strategies for the detection of differential expression using RNA-Sequencing

BackgroundRNA sequencing (RNA-Seq) has emerged as a powerful approach for the detection of differential gene expression with both high-throughput and high resolution capabilities possible depending upon the experimental design chosen. Multiplex experimental designs are now readily available, these can be utilised to increase the numbers of samples or replicates profiled at the cost of decreased sequencing depth generated per sample. These strategies impact on the power of the approach to accurately identify differential expression. This study presents a detailed analysis of the power to detect differential expression in a range of scenarios including simulated null and differential expression distributions with varying numbers of biological or technical replicates, sequencing depths and analysis methods.ResultsDifferential and non-differential expression datasets were simulated using a combination of negative binomial and exponential distributions derived from real RNA-Seq data. These datasets were used to evaluate the performance of three commonly used differential expression analysis algorithms and to quantify the changes in power with respect to true and false positive rates when simulating variations in sequencing depth, biological replication and multiplex experimental design choices.ConclusionsThis work quantitatively explores comparisons between contemporary analysis tools and experimental design choices for the detection of differential expression using RNA-Seq. We found that the DESeq algorithm performs more conservatively than edgeR and NBPSeq. With regard to testing of various experimental designs, this work strongly suggests that greater power is gained through the use of biological replicates relative to library (technical) replicates and sequencing depth. Strikingly, sequencing depth could be reduced as low as 15% without substantial impacts on false positive or true positive rates.

Analysis of Australian Crohn's disease pedigrees refines the localization for susceptibility to inflammatory bowel disease on chromosome 16.

A number of localizations for the putative susceptibility gene(s) have been identified for both Crohns disease and ulcerative colitis. In a genome wide scan, Hugot et al. (1996) identified a region on chromosome 16 which appeared to be responsible for the inheritance of inflammatory bowel disease in a small proportion of families. Subsequent work has suggested that this localization is important for susceptibility to Crohns disease rather than ulcerative colitis (Ohmen et al. 1996; Parkes et al. 1996). We investigated the contribution of this localization to the inheritance of inflammatory bowel disease in 54 multiplex Australian families, and confirmed its importance in a significant proportion of Crohns disease families; we further refined the localization to a region near to D16S409, obtaining a maximum LOD score of 6.3 between D16S409 and D16S753.

CARD15/NOD2 risk alleles in the development of Crohn's disease in the Australian population.

We have previously reported strong evidence for linkage between IBD1 and Crohns disease (CD) in Australian Crohns disease families. Three risk alleles for Crohns disease, (Arg702Trp (C/T), Gly908Arg (G/C) and 980fs981 (‐/C), were recently identified in the CARD15/NOD2 gene on chromosome 16, implicating this as the IBD1 locus. Using a novel diagnostic PCR‐RFLP, we have examined the frequency of these alleles in 205 multiplex IBD families, 107 sporadic Crohns disease cases and 409 normal individuals. We demonstrate that the three risk alleles are more frequent in Crohns disease, than in controls, with allelic frequencies of 0.11, 0.02 and 0.07 respectively. Heterozygosity for individual variants conferred a three‐fold increase in risk for Crohns disease while substantially higher risks were associated with being homozygous or compound heterozygous. Despite a significantly lower population allele frequency for the frameshift mutation than reported by other groups, we see a similar contribution by this allele to the risk of developing Crohns disease. While the three risk alleles influence susceptibility to Crohns disease in Australia, we show that these alleles do not fully explain the linkage evidence and suggest that there are very likely additional IBD1 susceptibility alleles yet to be described in Australian CD at the NOD2 locus. We also show a second linkage peak in Australian CD that provides some support for a second disease susceptibility locus on chromosome 16.

Plasma free metanephrines are superior to urine and plasma catecholamines and urine catecholamine metabolites for the investigation of phaeochromocytoma

Aim: To compare the relative diagnostic efficacy of several different tests used to establish a diagnosis of phaeochromocytoma, in patients with a proven diagnosis of phaeochromocytoma, and in hospital patients with significant disease of other types. Methods: We prospectively compared biochemical markers of catecholamine output and metabolism in plasma and urine in 22 patients with histologically proven phaeochromocytoma, 15 intensive care unit (ICU) patients, 30 patients on chronic haemodialysis and both hypertensive (n = 10) and normotensive (n = 16) controls. Results: Receiver operating characteristic curves were plotted. At the point of maximum efficiency, plasma free metanephrines showed 100% sensitivity and 97.6% specificity, compared with plasma catecholamines (78.6% and 70.7%), urine catecholamines (78.6% and 87.8%), urine metanephrines (85.7% and 95.1%), and urine hydroxymethoxymandelic acid (HMMA or VMA) (93.0% and 75.8%). All patients with phaeochromocytoma had plasma free metanephrine concentrations at least 27% above the upper limit of the reference range. Only three other patients (two on haemodialysis and one in ICU) had PFM concentrations more than 50% above the upper limit of the reference range. Conclusions: In patients with phaeochromocytoma, plasma free metanephrines displayed superior diagnostic sensitivity and specificity compared with other biochemical markers of catecholamine output and metabolism.

Statistical Analysis of Adsorption Models for Oligonucleotide Microarrays

Recent analyses have shown that the relationship between intensity measurements from high density oligonucleotide microarrays and known concentration is non linear. Thus many measurements of so-called gene expression are neither measures of transcript nor mRNA concentration as might be expected. Intensity as measured in such microarrays is a measurement of fluorescent dye attached to probe-target duplexes formed during hybridization of a sample to the probes on the microarray. We develop several dynamic adsorption models relating fluorescent dye intensity to target RNA concentration, the simplest of which is the equilibrium Langmuir isotherm, or hyperbolic response function. Using data from the Affymerix HG-U95A Latin Square experiment, we evaluate various physical models, including equilibrium and non-equilibrium models, by applying maximum likelihood methods. We show that for these data, equilibrium Langmuir isotherms with probe dependent parameters are appropriate. We describe how probe sequence information may then be used to estimate the parameters of the Langmuir isotherm in order to provide an improved measure of absolute target concentration.

The correlation between relatives under the multifactorial model with assortative mating

1 .

Discovery of serum biomarkers for pancreatic adenocarcinoma using proteomic analysis

Background and aims:The serum/plasma proteome was explored for biomarkers to improve the diagnostic ability of CA19-9 in pancreatic adenocarcinoma (PC).Methods:A Training Set of serum samples from 20 resectable and 18 stage IV PC patients, 54 disease controls (DCs) and 68 healthy volunteers (HVs) were analysed by surface-enhanced laser desorption and ionisation time-of-flight mass spectrometry (SELDI-TOF MS). The resulting protein panel was validated on 40 resectable PC, 21 DC and 19 HV plasma samples (Validation-1 Set) and further by ELISA on 33 resectable PC, 28 DC and 18 HV serum samples (Validation-2 Set). Diagnostic panels were derived using binary logistic regression incorporating internal cross-validation followed by receiver operating characteristic (ROC) analysis.Results:A seven-protein panel from the training set PC vs DC and from PC vs HV samples gave the ROC area under the curve (AUC) of 0.90 and 0.90 compared with 0.87 and 0.91 for CA19-9. The AUC was greater (0.97 and 0.99, P<0.05) when CA19-9 was added to the panels and confirmed on the validation-1 samples. A simplified panel of apolipoprotein C-I (ApoC-I), apolipoprotein A-II (ApoA-II) and CA19-9 was tested on the validation-2 set by ELISA, in which the ROC AUC was greater than that of CA19-9 alone for PC vs DC (0.90 vs 0.84) and for PC vs HV (0.96 vs 0.90).Conclusions:A simplified diagnostic panel of CA19-9, ApoC-I and ApoA-II improves the diagnostic ability of CA19-9 alone and may have clinical utility.

Impairment of organ-specific T cell negative selection by diabetes susceptibility genes: genomic analysis by mRNA profiling

BackgroundT cells in the thymus undergo opposing positive and negative selection processes so that the only T cells entering circulation are those bearing a T cell receptor (TCR) with a low affinity for self. The mechanism differentiating negative from positive selection is poorly understood, despite the fact that inherited defects in negative selection underlie organ-specific autoimmune disease in AIRE-deficient people and the non-obese diabetic (NOD) mouse strainResultsHere we use homogeneous populations of T cells undergoing either positive or negative selection in vivo together with genome-wide transcription profiling on microarrays to identify the gene expression differences underlying negative selection to an Aire-dependent organ-specific antigen, including the upregulation of a genomic cluster in the cytogenetic band 2F. Analysis of defective negative selection in the autoimmune-prone NOD strain demonstrates a global impairment in the induction of the negative selection response gene set, but little difference in positive selection response genes. Combining expression differences with genetic linkage data, we identify differentially expressed candidate genes, including Bim, Bnip3, Smox, Pdrg1, Id1, Pdcd1, Ly6c, Pdia3, Trim30 and Trim12.ConclusionThe data provide a molecular map of the negative selection response in vivo and, by analysis of deviations from this pathway in the autoimmune susceptible NOD strain, suggest that susceptibility arises from small expression differences in genes acting at multiple points in the pathway between the TCR and cell death.

Adsorption models of hybridization and post-hybridization behaviour on oligonucleotide microarrays

Analysis of data from an Affymetrix Latin Square spike-in experiment indicates that measured fluorescence intensities of features on an oligonucleotide microarray are related to spike-in RNA target concentrations via a hyperbolic response function, generally identified as a Langmuir adsorption isotherm. Furthermore, the asymptotic signal at high spike-in concentrations is almost invariably lower for a mismatch feature than for its partner perfect match feature. We survey a number of theoretical adsorption models of hybridization at the microarray surface and find that in general they are unable to explain the differing saturation responses of perfect and mismatch features. On the other hand, we find that a simple and consistent explanation can be found in a model in which equilibrium hybridization is followed by partial dissociation of duplexes during the post-hybridization washing phase.