M. Luz Calle

Comprehensive Transcriptional Analysis of Early-Stage Urothelial Carcinoma

Non-muscle-invasive bladder cancer (NMIBC) is a heterogeneous disease with widely different outcomes. We performed a comprehensive transcriptional analysis of 460 early-stage urothelial carcinomas and showed that NMIBC can be subgrouped into three major classes with basal- and luminal-like characteristics and different clinical outcomes. Large differences in biological processes such as the cell cycle, epithelial-mesenchymal transition, and differentiation were observed. Analysis of transcript variants revealed frequent mutations in genes encoding proteins involved in chromatin organization and cytoskeletal functions. Furthermore, mutations in well-known cancer driver genes (e.g., TP53 and ERBB2) were primarily found in high-risk tumors, together with APOBEC-related mutational signatures. The identification of subclasses in NMIBC may offer better prognostication and treatment selection based on subclass assignment.

Letter to the Editor: Stability of Random Forest importance measures

The goal of this article (letter to the editor) is to emphasize the value of exploring ranking stability when using the importance measures, mean decrease accuracy (MDA) and mean decrease Gini (MDG), provided by Random Forest. We illustrate with a real and a simulated example that ranks based on the MDA are unstable to small perturbations of the dataset and ranks based on the MDG provide more robust results.

Tutorial on methods for interval-censored data and their implementation in R

Interval censoring is encountered in many practical situations when the event of interest cannot be observed and it is only known to have occurred within a time window. The theory for the analysis of interval-censored data has been developed over the past three decades and several reviews have been written. However, it is still a common practice in medical and reliability studies to simplify the interval censoring structure of the data into a more standard right censoring situation by, for instance, imputing the midpoint of the censoring interval. The availability of software for right censoring might well be the main reason for this simplifying practice. In contrast, several methods have been developed to deal with interval-censored data and the corresponding algorithms to make the procedures feasible are scattered across the statistical software or remain behind the personal computers of many researchers. The purpose of this tutorial is to present, in a pedagogical and unified manner, the methodology and the available software for analyzing interval-censored data. The paper covers frequentist non-parametric, parametric and semiparametric estimating approaches, non-parametric tests for comparing survival curves and a section on simulation of interval-censored data. The methods and the software are described using the data from a dental study.

Model-Based Multifactor Dimensionality Reduction for detecting epistasis in case–control data in the presence of noise

Analyzing the combined effects of genes and/or environmental factors on the development of complex diseases is a great challenge from both the statistical and computational perspective, even using a relatively small number of genetic and nongenetic exposures. Several data‐mining methods have been proposed for interaction analysis, among them, the Multifactor Dimensionality Reduction Method (MDR) has proven its utility in a variety of theoretical and practical settings. Model‐Based Multifactor Dimensionality Reduction (MB‐MDR), a relatively new MDR‐based technique that is able to unify the best of both nonparametric and parametric worlds, was developed to address some of the remaining concerns that go along with an MDR analysis. These include the restriction to univariate, dichotomous traits, the absence of flexible ways to adjust for lower order effects and important confounders, and the difficulty in highlighting epistatic effects when too many multilocus genotype cells are pooled into two new genotype groups. We investigate the empirical power of MB‐MDR to detect gene–gene interactions in the absence of any noise and in the presence of genotyping error, missing data, phenocopy, and genetic heterogeneity. Power is generally higher for MB‐MDR than for MDR, in particular in the presence of genetic heterogeneity, phenocopy, or low minor allele frequencies.

mbmdr: an R package for exploring gene–gene interactions associated with binary or quantitative traits

SUMMARY We describe mbmdr, an R package for implementing the model-based multifactor dimensionality reduction (MB-MDR) method. MB-MDR has been proposed by Calle et al. as a dimension reduction method for exploring gene-gene interactions in case-control association studies. It is an extension of the popular multifactor dimensionality reduction (MDR) method of Ritchie et al. allowing a more flexible definition of risk cells. In MB-MDR, risk categories are defined using a regression model which allows adjustment for covariates and main effects and, in addition to the classical low risk and high risk categories, MB-MDR considers a third category of indeterminate or not informative cells. An important improvement added to the current mbmdr algorithm with respect to the original MB-MDR formulation in Calle et al. and also to the classical MDR approach, is the extension of the methodology to different outcome types. While MB-MDR was initially proposed for binary traits in the context of case-control studies, the mbmdr package provides options to analyze both binary or quantitative traits for unrelated individuals. AVAILABILITY http://cran.r-project.org/.

Genetic Susceptibility to Distinct Bladder Cancer Subphenotypes

BACKGROUND Clinical, pathologic, and molecular evidence indicate that bladder cancer is heterogeneous with pathologic/molecular features that define distinct subphenotypes with different prognoses. It is conceivable that specific patterns of genetic susceptibility are associated with particular subphenotypes. OBJECTIVE To examine evidence for the contribution of germline genetic variation to bladder cancer heterogeneity. DESIGN, SETTING, AND PARTICIPANTS The Spanish Bladder Cancer/EPICURO Study is a case-control study based in 18 hospitals located in five areas in Spain. Cases were patients with a newly diagnosed, histologically confirmed, urothelial cell carcinoma of the bladder from 1998 to 2001. Case diagnoses were reviewed and uniformly classified by pathologists following the World Health Organisation/International Society of Urological Pathology 1999 criteria. Controls were hospital-matched patients (n=1149). MEASUREMENTS A total of 1526 candidate variants in 423 candidate genes were analysed. Three distinct subphenotypes were defined according to stage and grade: low-grade nonmuscle invasive (n=586), high-grade nonmuscle invasive (n=219), and muscle invasive (n=246). The association between each variant and subphenotype was assessed by polytomous risk models adjusting for potential confounders. Heterogeneity in genetic susceptibility among subphenotypes was also tested. RESULTS AND LIMITATIONS Two established bladder cancer susceptibility genotypes, NAT2 slow-acetylation and GSTM1-null, exhibited similar associations among the subphenotypes, as did VEGF-rs25648, which was previously identified in our study. Other variants conferred risks for specific tumour subphenotypes such as PMS2-rs6463524 and CD4-rs3213427 (respective heterogeneity p values of 0.006 and 0.004), which were associated with muscle-invasive tumours (per-allele odds ratios [95% confidence interval] of 0.56 [0.41-0.77] and 0.71 [0.57-0.88], respectively) but not with non-muscle-invasive tumours. Heterogeneity p values were not robust in multiple testing according to their false-discovery rate. CONCLUSIONS These exploratory analyses suggest that genetic susceptibility loci might be related to the molecular/pathologic diversity of bladder cancer. Validation through large-scale replication studies and the study of additional genes and single nucleotide polymorphisms are required.

AUC-RF: A New Strategy for Genomic Profiling with Random Forest

Objective: Genomic profiling, the use of genetic variants at multiple loci simultaneously for the prediction of disease risk, requires the selection of a set of genetic variants that best predicts disease status. The goal of this work was to provide a new selection algorithm for genomic profiling. Methods: We propose a new algorithm for genomic profiling based on optimizing the area under the receiver operating characteristic curve (AUC) of the random forest (RF). The proposed strategy implements a backward elimination process based on the initial ranking of variables. Results and Conclusions: We demonstrate the advantage of using the AUC instead of the classification error as a measure of predictive accuracy of RF. In particular, we show that the use of the classification error is especially inappropriate when dealing with unbalanced data sets. The new procedure for variable selection and prediction, namely AUC-RF, is illustrated with data from a bladder cancer study and also with simulated data. The algorithm is publicly available as an R package, named AUCRF, at http://cran.r-project.org/.

FAM-MDR: a flexible family-based multifactor dimensionality reduction technique to detect epistasis using related individuals.

We propose a novel multifactor dimensionality reduction method for epistasis detection in small or extended pedigrees, FAM-MDR. It combines features of the Genome-wide Rapid Association using Mixed Model And Regression approach (GRAMMAR) with Model-Based MDR (MB-MDR). We focus on continuous traits, although the method is general and can be used for outcomes of any type, including binary and censored traits. When comparing FAM-MDR with Pedigree-based Generalized MDR (PGMDR), which is a generalization of Multifactor Dimensionality Reduction (MDR) to continuous traits and related individuals, FAM-MDR was found to outperform PGMDR in terms of power, in most of the considered simulated scenarios. Additional simulations revealed that PGMDR does not appropriately deal with multiple testing and consequently gives rise to overly optimistic results. FAM-MDR adequately deals with multiple testing in epistasis screens and is in contrast rather conservative, by construction. Furthermore, simulations show that correcting for lower order (main) effects is of utmost importance when claiming epistasis. As Type 2 Diabetes Mellitus (T2DM) is a complex phenotype likely influenced by gene-gene interactions, we applied FAM-MDR to examine data on glucose area-under-the-curve (GAUC), an endophenotype of T2DM for which multiple independent genetic associations have been observed, in the Amish Family Diabetes Study (AFDS). This application reveals that FAM-MDR makes more efficient use of the available data than PGMDR and can deal with multi-generational pedigrees more easily. In conclusion, we have validated FAM-MDR and compared it to PGMDR, the current state-of-the-art MDR method for family data, using both simulations and a practical dataset. FAM-MDR is found to outperform PGMDR in that it handles the multiple testing issue more correctly, has increased power, and efficiently uses all available information.

Frequentist and Bayesian approaches for interval-censored data

Interval censoring appears when the event of interest is only known to have occurred within a random time interval. Estimation and hypothesis testing procedures for interval-censored data are surveyed. We distinguish between frequentist and Bayesian approaches. Computational aspects for every proposed method are described and solutions with S-Plus, whenever are feasible, are mentioned. Three real data sets are analyzed.

Application of multi-SNP approaches Bayesian LASSO and AUC-RF to detect main effects of inflammatory-gene variants associated with bladder cancer risk

The relationship between inflammation and cancer is well established in several tumor types, including bladder cancer. We performed an association study between 886 inflammatory-gene variants and bladder cancer risk in 1,047 cases and 988 controls from the Spanish Bladder Cancer (SBC)/EPICURO Study. A preliminary exploration with the widely used univariate logistic regression approach did not identify any significant SNP after correcting for multiple testing. We further applied two more comprehensive methods to capture the complexity of bladder cancer genetic susceptibility: Bayesian Threshold LASSO (BTL), a regularized regression method, and AUC-Random Forest, a machine-learning algorithm. Both approaches explore the joint effect of markers. BTL analysis identified a signature of 37 SNPs in 34 genes showing an association with bladder cancer. AUC-RF detected an optimal predictive subset of 56 SNPs. 13 SNPs were identified by both methods in the total population. Using resources from the Texas Bladder Cancer study we were able to replicate 30% of the SNPs assessed. The associations between inflammatory SNPs and bladder cancer were reexamined among non-smokers to eliminate the effect of tobacco, one of the strongest and most prevalent environmental risk factor for this tumor. A 9 SNP-signature was detected by BTL. Here we report, for the first time, a set of SNP in inflammatory genes jointly associated with bladder cancer risk. These results highlight the importance of the complex structure of genetic susceptibility associated with cancer risk.