التفاصيل البيبلوغرافية
| العنوان: |
Exploration of isoform switching and mutation expression in breast cancer by mRNA-sequencing analysis. |
| المؤلفون: |
Hoadley, K. A., Parker, J. S., Wilkerson, M. D., Mose, L. E., Jefferys, S. R., Soloway, M. G., Turman, Y. J., Auman, J. T., Hayes, D. N., Perou, C. M. |
| المصدر: |
Cancer Research. Dec2012 Meeting Abstracts, Vol. 72 Issue 24a, p2231-2231. 1p. |
| مصطلحات موضوعية: |
*BREAST cancer research, *GENE expression, *CANCER, *DNA microarrays |
| مستخلص: |
Background: RNA expression, as well as other high dimensional assays, has helped delineate the complexity and diversity of breast cancer . Gene expression based classifications have identified at least five intrinsic subtypes that have distinct expression patterns and clinical outcomes. Here we explore mRNA-sequencing, accomplished using Next Generation Sequencing (NGS) technology , which provides additional information beyond that obtained from DNA microarray-based expression studies. Methods: As part of The Cancer Genome Atlas (TCGA) project, paired-end 2x50bp mRNA-sequencing was performed on over 700 breast tumors and 100 adjacent normal samples using an Illumina HiSeq2000. Sequences were aligned to the genome using MapSplice and gene and transcript abundance levels were determined using RSEM. Whole exome sequencing was also performed using the tumor and normal genomic DNA from each patient. Supervised identification of isoform switching (i.e. changing from one gene isoform to another) was performed by identifying isoforms that had significant t-statistics of opposing sign. Hierarchical clustering and isoform-guided supervised analysis were used to characterize patterns of isoform switching and their associated expression correlates. Results: Unsupervised and supervised hierarchical clustering analysis of mRNA-seq data identifies tumor intrinsic subtypes and all of the common expression features previously seen in DNA microarray studies. The 582 breast cancers processed thus far were classified as 43% luminal A, 23% luminal B, 14% basal-like, 9% Her2-enriched, and 11% normal-like. Isoform level switching analyses were performed using supervised and unsupervised methods. A bias in isoform quantitation was first identified that correlated with RNA Integrity Score (i.e. RIN). When this bias was accounted for , unsupervised methods identified 9 genes that demonstrated significant isoform switching that was independent of subtype. Supervised analysis of basal-like versus luminal tumors identified 30 genes showing isoform switching including CD44, CTNND1 and RAD51L1; most of these unique isoform switches also correlated with the expression of many other genes that did not show isoform changes. Additional analyses are being performed including 1) identification of genes expressing more than one isoform, but which do not change dramatically in abundance across samples, and 2) correlations of genomic DNA somatic mutations with the mRNA-seq based mutation detection. Conclusions: Appropriate processing of mRNA-sequencing data yields unparalleled information in isoform level abundance and robust expression signatures. The modest isoform switching observed here segregated samples into distinct groups of common switching events, most of which correlated with subtype. The additional data types provided by mRNA-seq including isoform identification, fusion gene detection, and mutation information, extend the capabilities of expression analyses and may provide new data of clinical utility. [ABSTRACT FROM AUTHOR] |
| قاعدة البيانات: |
Academic Search Index |
Full Text Finder