Dataset Information

Observed Survival Interval: A Supplement to TCGA Pan-Cancer Clinical Data Resource.

ABSTRACT: To drive high-quality omics translational research using The Cancer Genome Atlas (TCGA) data, a TCGA Pan-Cancer Clinical Data Resource was proposed. However, there is an out-of-step issue between clinical outcomes and the omics data of TCGA for skin cutaneous melanoma (SKCM), due to the majority of metastatic samples. In clinical cases, the survival time started from the initial SKCM diagnosis, while the omics data were characterized at TCGA sampling. This study aimed to address this issue by proposing an observed survival interval (OBS), which was defined as the time interval from TCGA sampling to patient death or last follow-up. We compared the OBS with the usual recommended overall survival (OS) by associating them with both clinical data and microRNA sequencing data of TCGA-SKCM. We found that the OS of primary SKCM was significantly shorter than that of metastatic SKCM, while the opposite happened if OBS was compared. OS was associated with the pathological stage of both primary and metastatic SKCM, while OBS was associated with the pathological stage of primary SKCM but not that of metastatic SKCM. Five previously cross-validated survival-associated microRNAs were found to be associated with the OBS rather than OS in metastatic SKCM. Thus, the OBS was more appropriate for associating microRNA-omics data of TCGA-SKCM than OS, and it is a timely supplement to TCGA Pan-Cancer Clinical Data Resource.

SUBMITTER: Xiong J

PROVIDER: S-EPMC6468755 | biostudies-literature | 2019 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Observed Survival Interval: A Supplement to TCGA Pan-Cancer Clinical Data Resource.

Xiong Jie J Bing Zhitong Z Guo Shengyu S

Cancers 20190226 3

To drive high-quality omics translational research using The Cancer Genome Atlas (TCGA) data, a TCGA Pan-Cancer Clinical Data Resource was proposed. However, there is an out-of-step issue between clinical outcomes and the omics data of TCGA for skin cutaneous melanoma (SKCM), due to the majority of metastatic samples. In clinical cases, the survival time started from the initial SKCM diagnosis, while the omics data were characterized at TCGA sampling. This study aimed to address this issue by pr ...[more]

PMID: 30813652

Similar Datasets

Project description:Quantitative PCR (qPCR) remains the most widely used technique for gene expression evaluation. Obtaining reliable data using this method requires reference genes (RGs) with stable mRNA level under experimental conditions. This issue is especially crucial in cancer studies because each tumor has a unique molecular portrait. The Cancer Genome Atlas (TCGA) project provides RNA-Seq data for thousands of samples corresponding to dozens of cancers and presents the basis for assessment of the suitability of genes as reference ones for qPCR data normalization. Using TCGA RNA-Seq data and previously developed CrossHub tool, we evaluated mRNA level of 32 traditionally used RGs in 12 cancer types, including those of lung, breast, prostate, kidney, and colon. We developed an 11-component scoring system for the assessment of gene expression stability. Among the 32 genes, PUM1 was one of the most stably expressed in the majority of examined cancers, whereas GAPDH, which is widely used as a RG, showed significant mRNA level alterations in more than a half of cases. For each of 12 cancer types, we suggested a pair of genes that are the most suitable for use as reference ones. These genes are characterized by high expression stability and absence of correlation between their mRNA levels. Next, the scoring system was expanded with several features of a gene: mutation rate, number of transcript isoforms and pseudogenes, participation in cancer-related processes on the basis of Gene Ontology, and mentions in PubMed-indexed articles. All the genes covered by RNA-Seq data in TCGA were analyzed using the expanded scoring system that allowed us to reveal novel promising RGs for each examined cancer type and identify several "universal" pan-cancer RG candidates, including SF3A1, CIAO1, and SFRS4. The choice of RGs is the basis for precise gene expression evaluation by qPCR. Here, we suggested optimal pairs of traditionally used RGs for 12 cancer types and identified novel promising RGs that demonstrate high expression stability and other features of reliable and convenient RGs (high expression level, low mutation rate, non-involvement in cancer-related processes, single transcript isoform, and absence of pseudogenes).

Project description:BackgroundPan-omics, pan-cancer analysis has advanced our understanding of the molecular heterogeneity of cancer. However, such analyses have been limited in their ability to use information from multiple sources of data (e.g., omics platforms) and multiple sample sets (e.g., cancer types) to predict clinical outcomes. We address the issue of prediction across multiple high-dimensional sources of data and sample sets by using molecular patterns identified by BIDIFAC+, a method for integrative dimension reduction of bidimensionally-linked matrices, in a Bayesian hierarchical model. Our model performs variable selection through spike-and-slab priors that borrow information across clustered data. We use this model to predict overall patient survival from the Cancer Genome Atlas with data from 29 cancer types and 4 omics sources and use simulations to characterize the performance of the hierarchical spike-and-slab prior.ResultsWe found that molecular patterns shared across all or most cancers were largely not predictive of survival. However, our model selected patterns unique to subsets of cancers that differentiate clinical tumor subtypes with markedly different survival outcomes. Some of these subtypes were previously established, such as subtypes of uterine corpus endometrial carcinoma, while others may be novel, such as subtypes within a set of kidney carcinomas. Through simulations, we found that the hierarchical spike-and-slab prior performs best in terms of variable selection accuracy and predictive power when borrowing information is advantageous, but also offers competitive performance when it is not.ConclusionsWe address the issue of prediction across multiple sources of data by using results from BIDIFAC+ in a Bayesian hierarchical model for overall patient survival. By incorporating spike-and-slab priors that borrow information across cancers, we identified molecular patterns that distinguish clinical tumor subtypes within a single cancer and within a group of cancers. We also corroborate the flexibility and performance of using spike-and-slab priors as a Bayesian variable selection approach.

Dataset Information

Observed Survival Interval: A Supplement to TCGA Pan-Cancer Clinical Data Resource.

Publications

Observed Survival Interval: A Supplement to TCGA Pan-Cancer Clinical Data Resource.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets