Integration of survival data from multiple studies

We introduce a statistical procedure that integrates datasets from multiple biomedical studies to predict patients' survival, based on individual clinical and genomic profiles. The proposed procedure accounts for potential differences in the relation between predictors and outcomes across studies, due to distinct patient populations, treatments and technologies to measure outcomes and biomarkers. These differences are modeled explicitly with study-specific parameters. We use hierarchical regularization to shrink the study-specific parameters towards each other and to borrow information across studies. The estimation of the study-specific parameters utilizes a similarity matrix, which summarizes differences and similarities of the relations between covariates and outcomes across studies. We illustrate the method in a simulation study and using a collection of gene expression datasets in ovarian cancer. We show that the proposed model increases the accuracy of survival predictions compared to alternative meta-analytic methods.

PREDICTION OF HEREDITARY CANCERS USING NEURAL NETWORKS

Family history is a major risk factor for many types of cancer. Mendelian risk prediction models translate family histories into cancer risk predictions, based on knowledge of cancer susceptibility genes. These models are widely used in clinical practice to help identify high-risk individuals. Mendelian models leverage the entire family history, but they rely on many assumptions about cancer susceptibility genes that are either unrealistic or challenging to validate, due to low mutation prevalence. Training more flexible models, such as neural networks, on large databases of pedigrees can potentially lead to accuracy gains. In this paper we develop a framework to apply neural networks to family history data and investigate their ability to learn inherited susceptibility to cancer. While there is an extensive literature on neural networks and their state-of-the-art performance in many tasks, there is little work applying them to family history data. We propose adaptations of fully-connected neural networks and convolutional neural networks to pedigrees. In data simulated under Mendelian inheritance, we demonstrate that our proposed neural network models are able to achieve nearly optimal prediction performance. Moreover, when the observed family history includes misreported cancer diagnoses, neural networks are able to outperform the Mendelian BRCAPRO model embedding the correct inheritance laws. Using a large dataset of over 200,000 family histories, the Risk Service cohort, we train prediction models for future risk of breast cancer. We validate the models using data from the Cancer Genetics Network.

Quantitative Cluster Headache Analysis for Neurological Diagnosis Support Using Statistical Classification

Cluster headache (CH) belongs to the group III of The International Classification of Headaches. It is characterized by attacks of severe pain in the ocular/periocular area accompanied by cranial autonomic signs, including parasympathetic activation and sympathetic hypofunction on the symptomatic side. Iris pigmentation occurs in the neonatal period and depends on the sympathetic tone in each eye. We hypothesized that the presence of visible or subtle color iris changes in both eyes could be used as a quantitative biomarker for screening and early detection of CH. This work scrutinizes the scope of an automatic diagnosis-support system for early detection of CH, by using as indicator the error rate provided by a statistical classifier designed to identify the eye (left vs. right) from iris pixels in color images. Systematic tests were performed on a database with images of 11 subjects (four with CH, four with other ophthalmic diseases affecting the iris pigmentation, and three control subjects). Several aspects were addressed to design the classifier, including: (a) the most convenient color space for the statistical classifier; (b) whether the use of features associated to several color spaces is convenient; (c) the robustness of the classifier to iris spatial subregions; (d) the contribution of the pixels neighborhood. Our results showed that a reduced value for the error rate (lower than 0.25) can be used as CH marker, whereas structural regions of the iris image need to be taken into account. The iris color feature analysis using statistical classification is a potentially useful technique to investigate disorders affecting the autonomous nervous system in CH.

Early adolescent brain markers of late adolescent academic functioning

Academic performance in adolescence strongly influences adult prospects. Intelligence quotient (IQ) has historically been considered a strong predictor of academic performance. Less objectively explored have been morphometric features. We analyzed brain MRI morphometry metrics in early adolescence (age 12-14 years) as quantitative predictors of academic performance over high school using a naïve Bayesian classifier approach with n = 170 subjects. Based on the mean GPA, subjects were divided into high (GPA ≥3.54; n = 87) and low (GPA <3.54; n = 83) academic performers. Covariance analysis was performed to look at the influence of subject demographics. We examined predictive features from the 343 available regions (surface areas, cortical thickness, and subcortical volumes) and applied 4 algorithms for selection and reduction of attributes using Weka. Cortical thickness measures performed better than surface areas or subcortical volumes as predictors of academic performance. We identified 15 cortical thickness regions most predictive of academic performance, three of which have not been described in the literature predictive of academic performance. These were in the left hemisphere fusiform, bilateral insula, and left hemisphere paracentral regions. Prediction had a sensitivity of 0.65 and specificity of 0.73 with independent validation. Follow-up independent t-test analyses between high and low academic achievers on 10 of 15 regions showed between-group significance at the p < 0.05 level. High achievers demonstrated thicker cortices than low achievers. These newly identified regions may help pinpoint new targets for further study in understanding the developing adolescent brain in the classroom setting.

Low-cost scalable discretization, prediction, and feature selection for complex systems

Finding reliable discrete approximations of complex systems is a key prerequisite when applying many of the most popular modeling tools. Common discretization approaches (e.g., the very popular K-means clustering) are crucially limited in terms of quality, parallelizability, and cost. We introduce a low-cost improved quality scalable probabilistic approximation (SPA) algorithm, allowing for simultaneous data-driven optimal discretization, feature selection, and prediction. We prove its optimality, parallel efficiency, and a linear scalability of iteration cost. Cross-validated applications of SPA to a range of large realistic data classification and prediction problems reveal marked cost and performance improvements. For example, SPA allows the data-driven next-day predictions of resimulated surface temperatures for Europe with the mean prediction error of 0.75°C on a common PC (being around 40% better in terms of errors and five to six orders of magnitude cheaper than with common computational instruments used by the weather services).

MetaGxData: Clinically Annotated Breast, Ovarian and Pancreatic Cancer Datasets and their Use in Generating a Multi-Cancer Gene Signature

A wealth of transcriptomic and clinical data on solid tumours are under-utilized due to unharmonized data storage and format. We have developed the MetaGxData package compendium, which includes manually-curated and standardized clinical, pathological, survival, and treatment metadata across breast, ovarian, and pancreatic cancer data. MetaGxData is the largest compendium of curated transcriptomic data for these cancer types to date, spanning 86 datasets and encompassing 15,249 samples. Open access to standardized metadata across cancer types promotes use of their transcriptomic and clinical data in a variety of cross-tumour analyses, including identification of common biomarkers, and assessing the validity of prognostic signatures. Here, we demonstrate that MetaGxData is a flexible framework that facilitates meta-analyses by using it to identify common prognostic genes in ovarian and breast cancer. Furthermore, we use the data compendium to create the first gene signature that is prognostic in a meta-analysis across 3 cancer types. These findings demonstrate the potential of MetaGxData to serve as an important resource in oncology research, and provide a foundation for future development of cancer-specific compendia.

Prediction of annual dengue incidence by hydro-climatic extremes for southern Taiwan

Dengue is one of the most rapidly spreading mosquito-borne viral diseases in the world. An increase in the incidence of dengue is commonly thought to be a consequence of variability of weather conditions. Taiwan, which straddles the Tropic of Cancer, is an excellent place to study the relationship between weather conditions and dengue fever cases since the island forms an isolated geographic environment. Therefore, clarifying the association between extreme weather conditions and annual dengue incidence is one of important issues for epidemic early warning. In this paper, we develop a Poisson regression model with extreme weather parameters for prediction of annual dengue incidence. A leave-one-out method is used to evaluate the performance of predicting dengue incidence. Our results indicate that dengue transmission has a positive relationship with the minimum temperature predictors during the early summer while a negative relationship with all the maximum 24-h rainfall predictors during the early epidemic phase of dengue outbreaks. Our findings provide a better understanding of the relationships between extreme weather and annual trends in dengue cases in Taiwan and it could have important implications for dengue forecasts in surrounding areas with similar meteorological conditions.

Training replicable predictors in multiple studies

This article considers replicability of the performance of predictors across studies. We suggest a general approach to investigating this issue, based on ensembles of prediction models trained on different studies. We quantify how the common practice of training on a single study accounts in part for the observed challenges in replicability of prediction performance. We also investigate whether ensembles of predictors trained on multiple studies can be combined, using unique criteria, to design robust ensemble learners trained upfront to incorporate replicability into different contexts and populations.