Dataset Information

Pathway-Informed Classification System (PICS) for Cancer Analysis Using Gene Expression Data.

ABSTRACT: We introduce Pathway-Informed Classification System (PICS) for classifying cancers based on tumor sample gene expression levels. PICS is a computational method capable of expeditiously elucidating both known and novel biological pathway involvement specific to various cancers and uses that learned pathway information to separate patients into distinct classes. The method clearly separates a pan-cancer dataset by tissue of origin and also sub-classifies individual cancer datasets into distinct survival classes. Gene expression values are collapsed into pathway scores that reveal which biological activities are most useful for clustering cancer cohorts into subtypes. Variants of the method allow it to be used on datasets that do and do not contain noncancerous samples. Activity levels of all types of pathways, broadly grouped into metabolic, cellular processes and signaling, and immune system, are useful for separating the pan-cancer cohort. In the clustering of specific cancer types, certain pathway types become more valuable depending on the site being studied. For lung cancer, signaling pathways dominate; for pancreatic cancer, signaling and metabolic pathways dominate; and for melanoma, immune system pathways are the most useful. This work suggests the utility of pathway-level genomic analysis and points in the direction of using pathway classification for predicting the efficacy and side effects of drugs and radiation.

SUBMITTER: Young MR

PROVIDER: S-EPMC4965015 | biostudies-literature | 2016

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Pathway-Informed Classification System (PICS) for Cancer Analysis Using Gene Expression Data.

Young Michael R MR Craft David L DL

Cancer informatics 20160727

We introduce Pathway-Informed Classification System (PICS) for classifying cancers based on tumor sample gene expression levels. PICS is a computational method capable of expeditiously elucidating both known and novel biological pathway involvement specific to various cancers and uses that learned pathway information to separate patients into distinct classes. The method clearly separates a pan-cancer dataset by tissue of origin and also sub-classifies individual cancer datasets into distinct su ...[more]

PMID: 27486299

Similar Datasets

Project description:BackgroundThe Cancer Genome Atlas (TCGA) has generated comprehensive molecular profiles. We aim to identify a set of genes whose expression patterns can distinguish diverse tumor types. Those features may serve as biomarkers for tumor diagnosis and drug development.MethodsUsing RNA-seq expression data, we undertook a pan-cancer classification of 9,096 TCGA tumor samples representing 31 tumor types. We randomly assigned 75% of samples into training and 25% into testing, proportionally allocating samples from each tumor type.ResultsWe could correctly classify more than 90% of the test set samples. Accuracies were high for all but three of the 31 tumor types, in particular, for READ (rectum adenocarcinoma) which was largely indistinguishable from COAD (colon adenocarcinoma). We also carried out pan-cancer classification, separately for males and females, on 23 sex non-specific tumor types (those unrelated to reproductive organs). Results from these gender-specific analyses largely recapitulated results when gender was ignored. Remarkably, more than 80% of the 100 most discriminative genes selected from each gender separately overlapped. Genes that were differentially expressed between genders included BNC1, FAT2, FOXA1, and HOXA11. FOXA1 has been shown to play a role for sexual dimorphism in liver cancer. The differentially discriminative genes we identified might be important for the gender differences in tumor incidence and survival.ConclusionsWe were able to identify many sets of 20 genes that could correctly classify more than 90% of the samples from 31 different tumor types using TCGA RNA-seq data. This accuracy is remarkable given the number of the tumor types and the total number of samples involved. We achieved similar results when we analyzed 23 non-sex-specific tumor types separately for males and females. We regard the frequency with which a gene appeared in those sets as measuring its importance for tumor classification. One third of the 50 most frequently appearing genes were pseudogenes; the degree of enrichment may be indicative of their importance in tumor classification. Lastly, we identified a few genes that might play a role in sexual dimorphism in certain cancers.

Project description:BACKGROUND: Designing appropriate machine learning methods for identifying genes that have a significant discriminating power for disease outcomes has become more and more important for our understanding of diseases at genomic level. Although many machine learning methods have been developed and applied to the area of microarray gene expression data analysis, the majority of them are based on linear models, which however are not necessarily appropriate for the underlying connection between the target disease and its associated explanatory genes. Linear model based methods usually also bring in false positive significant features more easily. Furthermore, linear model based algorithms often involve calculating the inverse of a matrix that is possibly singular when the number of potentially important genes is relatively large. This leads to problems of numerical instability. To overcome these limitations, a few non-linear methods have recently been introduced to the area. Many of the existing non-linear methods have a couple of critical problems, the model selection problem and the model parameter tuning problem, that remain unsolved or even untouched. In general, a unified framework that allows model parameters of both linear and non-linear models to be easily tuned is always preferred in real-world applications. Kernel-induced learning methods form a class of approaches that show promising potentials to achieve this goal. RESULTS: A hierarchical statistical model named kernel-imbedded Gaussian process (KIGP) is developed under a unified Bayesian framework for binary disease classification problems using microarray gene expression data. In particular, based on a probit regression setting, an adaptive algorithm with a cascading structure is designed to find the appropriate kernel, to discover the potentially significant genes, and to make the optimal class prediction accordingly. A Gibbs sampler is built as the core of the algorithm to make Bayesian inferences. Simulation studies showed that, even without any knowledge of the underlying generative model, the KIGP performed very close to the theoretical Bayesian bound not only in the case with a linear Bayesian classifier but also in the case with a very non-linear Bayesian classifier. This sheds light on its broader usability to microarray data analysis problems, especially to those that linear methods work awkwardly. The KIGP was also applied to four published microarray datasets, and the results showed that the KIGP performed better than or at least as well as any of the referred state-of-the-art methods did in all of these cases. CONCLUSION: Mathematically built on the kernel-induced feature space concept under a Bayesian framework, the KIGP method presented in this paper provides a unified machine learning approach to explore both the linear and the possibly non-linear underlying relationship between the target features of a given binary disease classification problem and the related explanatory gene expression data. More importantly, it incorporates the model parameter tuning into the framework. The model selection problem is addressed in the form of selecting a proper kernel type. The KIGP method also gives Bayesian probabilistic predictions for disease classification. These properties and features are beneficial to most real-world applications. The algorithm is naturally robust in numerical computation. The simulation studies and the published data studies demonstrated that the proposed KIGP performs satisfactorily and consistently.

Dataset Information

Pathway-Informed Classification System (PICS) for Cancer Analysis Using Gene Expression Data.

Publications

Pathway-Informed Classification System (PICS) for Cancer Analysis Using Gene Expression Data.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets