TCGASpliceSeq a compendium of alternative mRNA splicing in cancer.

Ryan M, Wong WC, Brown R, Akbani R, Su X, Broom B, Melott J, Weinstein J. · 2016

Read TCGASpliceSeq a compendium of alternative mRNA splicing in cancer. by Ryan M, Wong WC, Brown R, Akbani R, Su X, Broom B, Melott J, Weinstein J. (2016) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/europmc-pmc4702910.txt.

Introduction

Michael Ryan, Wing Chung Wong, Robert Brown, and five colleagues built TCGASpliceSeq, described in Nucleic Acids Research in 2016. Alternative splicing lets one gene yield several protein versions, and cancers often rewire those choices. The Cancer Genome Atlas holds RNA data from over ten thousand tumors across thirty-three types, a rich splicing record that was too heavy for most researchers to process. The team's SpliceSeq software computed splicing measures for every sample in advance.

The website serves those precomputed values through a visual interface. Users can inspect one gene, hunt for the strongest splicing differences between tumor types, or compare tumor with nearby normal tissue. Charts show splicing patterns, read counts, and statistics together, and data can be downloaded for bigger analyses. Access is free to academic, government, and commercial users alike. The resource turns an intimidating archive into something a bench scientist can query in minutes.

Catalog record

CREATORRyan M, Wong WC, Brown R, Akbani R, Su X, Broom B, Melott J, Weinstein J.
YEAR2016
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21 KB
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC4702910

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

TCGASpliceSeq a compendium of alternative mRNA splicing in cancer By Michael Ryan, Wing Chung Wong, Robert Brown, Rehan Akbani, Xiaoping Su, Bradley Broom, James Melott, John Weinstein Published in: Nucleic Acids Research DOI: 10.1093/nar/gkv1288 ABSTRACT TCGA's RNASeq data represent one of the largest collections of cancer transcriptomes ever assembled. RNASeq technology, combined with computational tools like our SpliceSeq package, provides a comprehensive, detailed view of alternative mRNA splicing. Aberrant splicing patterns in cancers have been implicated in such processes as carcinogenesis, de-differentiation and metastasis. TCGA SpliceSeq (http://bioinformatics.mdanderson.org/TCGASpliceSeq) is a web-based resource that provides a quick, user-friendly, highly visual interface for exploring the alternative splicing patterns of TCGA tumors. Percent Spliced In (PSI) values for splice events on samples from 33 different tumor types, including available adjacent normal samples, have been loaded into TCGA SpliceSeq. Investigators can interrogate genes of interest, search for the genes that show the strongest variation between or among selected tumor types, or explore splicing pattern changes between tumor and adjacent normal samples. The interface presents intuitive graphical representations of splicing patterns, read counts and various statistical summaries, including percent spliced in. Splicing data can also be downloaded for inclusion in integrative analyses. TCGA SpliceSeq is freely available for academic, government or commercial use. INTRODUCTION Alternative mRNA splicing is a regulated process that occurs in nearly all multi-exon human genes (1). The splicing mechanism enables a single gene to produce multiple protein products, and changes in splicing patterns are closely associated with development and differentiation. Of particular interest to cancer researchers are the alternate splicing changes in tumor cells that may represent a return to developmental splicing patterns of metabolism, angiogenesis, mobility and proliferation (2,3). The Cancer Genome Atlas (TCGA) project has generated a very large, public collection of multiple types of ‘omic’ data on samples from >10,000 cancer patients, across 33 different tumor types. The TCGA data provide a rich source for the investigation of alternative splicing patterns in cancer. In practice, however, integrating TCGA splicing data into broader analytical investigations is a prohibitively complex, compute-intensive process for most cancer genomics researchers. Exon, splice and transcript isoform expression levels are available from the TCGA Data Portal but do not come with: a unified gene model to support comparative analysis, splice event measurements, tumor specific summarizations, splice pattern graphical displays, protein impact information or simple query/download features for integrative analysis. To address these needs, we used SpliceSeq (4) to analyze thousands of TCGA RNASeq samples, computing and storing the consequent statistics in a TCGA splicing database. The splicing data are available through our TCGA SpliceSeq website. The site presents splicing patterns and related statistics in an intuitive, interactive, graphical form with dynamic features for exploration of cross-tumor-type or tumor-normal splice variations. TCGA SpliceSeq's data query/download capability enables researchers to retrieve per-sample, per-splice event measures of alternative splicing that can then be included in any type of integrative analysis. It also provides search functions that identify splice event variation across selected tumor types. Hence, it can be used to focus on particular genes, or, in knowledge discovery mode, to identify the most striking splice events. TCGA SpliceSeq is freely available for academic, government or commercial use at http://bioinformatics.mdanderson.org/TCGASpliceSeq/. DATABASE CONSTRUCTION We used our SpliceSeq tool to analyze the mRNA splicing patterns of TCGA samples. SpliceSeq starts with a reference model for each gene constructed from all of the gene's protein-coding transcripts in the Ensembl gene database (5). The transcript isoforms of each gene are assembled into a unified splice graph such that mRNA reads will align to a single unique location in the graph (Figure 1). TCGA sample reads are aligned to the splice graphs, and summary statistics are generated for each exon and splice junction. Read totals are normalized by exon length and the number of aligned reads in the sample (reads per 1000 bases of exon per million aligned reads). A second pass alignment is performed on unaligned reads looking for novel splices not present in the original Ensembl models. Novel exons are not identified by this algorithm. [Figure: SpliceSeq analysis of mRNA data. (A) Ensemble coding transcripts for each gene are assembled into a unified splice graph. (B) Reads are aligned to the splice graphs and totals for each exon and splice are calculated. Read counts are normalized to the length of each exon and number of aligned reads in the sample. (C) Percent Spliced In values are computed for each possible splice event in each gene. PSI is the ratio of reads indicating the presence of a transcript element versus the total reads covering the event. In this example, the 8 yellow reads (exon 3 body reads, exon 2–3 junction reads and exon 3–4 junction reads) indicate that exon 3 is present. The red junction 2–4 reads indicate that exon 3 was spliced out. The PSI is therefore 8/10 or 0.8 indicating that 80% of the transcripts in the sample include exon 3. (D) SpliceSeq evaluates all of the listed types of splice events.] The Percent Spliced In (PSI) value is a common, intuitive ratio for quantifying splicing events (1). The number of reads indicating that a transcript element is present is divided by the total number of reads covering the splicing event. In the example of an exon skip event in Figure 1C, a PSI value of 0.8 indicates that approximately 80% of the transcripts in the sample include exon 3 and 20% do not. PSI values were calculated for each potential splice event on all genes expressed in the TCGA samples. There are seven types of splice events: exon skip, alternate 5′ donor, alternate 3′ acceptor, retained intron, mutually exclusive exons, alternate first exon and alternate last exon. The results of SpliceSeq analysis of TCGA samples are stored in the TCGA SpliceSeq relational database. To date, we have fully loaded 15 tumor types, the 12 included in the first TCGA pan-can analysis and 3 additional tumor types. The fully loaded data contain PSI values for 80,000+ splice events on 19,036 genes for 6238 tumor samples and 496 adjacent normal samples. Aligning and loading the data is a computationally intensive task. For TCGA tissues that have not yet been fully loaded into the database, we analyzed 15 randomly selected tumor and 15 randomly selected adjacent normal samples (when available) to provide an initial overview of splicing in these tumor types. As additional full data sets are loaded, they will replace the sub-sampled data. Full data sets are indicated by a ‘***’ next to the tumor type on the PSI Data Download page. PSI values are not provided for a sample unless at least 8 reads cover the event (include + exclude reads ≥ 8). Novel splices are those that were not included in the reference Ensembl gene models but were detected in our second pass alignment. These can be spurious or very weakly expressed so if a splice event includes a novel splice, PSI values for the event will only be reported if the novel splice is expressed in ≥ 1% of transcripts of 10 or more samples. DATABASE USAGE The TCGA SpliceSeq database can be queried and explored via our graphical web application at http://bioinformatics.mdanderson.org/TCGASpliceSeq. Investigate a gene of interest The splicing

Read the complete text