Penalized Ordinal Regression Methods for Predicting Stage of Cancer in High-Dimensional Covariate Spaces

Quasi-Unimodal Distributions for Ordinal Classification

Ordinal classification tasks are present in a large number of different domains. However, common losses for deep neural networks, such as cross-entropy, do not properly weight the relative ordering between classes. For that reason, many losses have been proposed in the literature, which model the output probabilities as following a unimodal distribution. This manuscript reviews many of these losses on three different datasets and suggests a potential improvement that focuses the unimodal constraint on the neighborhood around the true class, allowing for a more flexible distribution, aptly called quasi-unimodal loss. For this purpose, two constraints are proposed: A first constraint concerns the relative order of the top-three probabilities, and a second constraint ensures that the remaining output probabilities are not higher than the top three. Therefore, gradient descent focuses on improving the decision boundary around the true class in detriment to the more distant classes. The proposed loss is found to be competitive in several cases.

Race/ethnicity-associated blood DNA methylation differences between Japanese and European American women: an exploratory study

BACKGROUND

Racial/ethnic disparities in health reflect a combination of genetic and environmental causes, and DNA methylation may be an important mediator. We compared in an exploratory manner the blood DNA methylome of Japanese Americans (JPA) versus European Americans (EUA).

METHODS

Genome-wide buffy coat DNA methylation was profiled among healthy Multiethnic Cohort participant women who were Japanese (JPA; n = 30) or European (EUA; n = 28) Americans aged 60-65. Differentially methylated CpGs by race/ethnicity (DM-CpGs) were identified by linear regression (Bonferroni-corrected P < 0.1) and analyzed in relation to corresponding gene expression, a priori selected single nucleotide polymorphisms (SNPs), and blood biomarkers of inflammation and metabolism using Pearson or Spearman correlations (FDR < 0.1).

RESULTS

We identified 174 DM-CpGs with the majority of hypermethylated in JPA compared to EUA (n = 133), often in promoter regions (n = 48). Half (51%) of the genes corresponding to the DM-CpGs were involved in liver function and liver disease, and the methylation in nine genes was significantly correlated with gene expression for DM-CpGs. A total of 156 DM-CpGs were associated with rs7489665 (SH2B1). Methylation of DM-CpGs was correlated with blood levels of the cytokine MIP1B (n = 146). We confirmed some of the DM-CpGs in the TCGA adjacent non-tumor liver tissue of Asians versus EUA.

CONCLUSION

We found a number of differentially methylated CpGs in blood DNA between JPA and EUA women with a potential link to liver disease, specific SNPs, and systemic inflammation. These findings may support further research on the role of DNA methylation in mediating some of the higher risk of liver disease among JPA.

Penalized negative binomial models for modeling an overdispersed count outcome with a high-dimensional predictor space: Application predicting micronuclei frequency

Chromosomal aberrations, such as micronuclei (MN), have served as biomarkers of genotoxic exposure and cancer risk. Guidelines for the process of scoring MN have been presented by the HUman MicroNucleus (HUMN) project. However, these guidelines were developed for assay performance but do not address how to statistically analyze the data generated by the assay. This has led to the application of various statistical methods that may render different interpretations and conclusions. By combining MN with data from other high-throughput genomic technologies such as gene expression microarray data, we may elucidate molecular features involved in micronucleation. Traditional methods that can model discrete (synonymously, count) data, such as MN frequency, require that the number of explanatory variables (P) is less than the number of samples (N). Due to this limitation, penalized models have been developed to enable model fitting for such over-parameterized datasets. Because penalized models in the discrete response setting are lacking, particularly when the count outcome is over-dispersed, herein we present our penalized negative binomial regression model that can be fit when P > N. Using simulation studies we demonstrate the performance of our method in comparison to commonly used penalized Poisson models when the outcome is over-dispersed and applied it to MN frequency and gene expression data collected as part of the Norwegian Mother and Child Cohort Study. Our countgmifs R package is available for download from the Comprehensive R Archive Network and can be applied to datasets having a discrete outcome that is either Poisson or negative binomial distributed and a high-dimensional covariate space.