Person:

Beck, Andrew

Gene set enrichment analysis (GSEA) associates gene sets and phenotypes, its use is predicated on the choice of a pre-defined collection of sets. The defacto standard implementation of GSEA provides seven collections yet there are no guidelines for the choice of collections and the impact of such choice, if any, is unknown. Here we compare each of the standard gene set collections in the context of a large dataset of drug response in human cancer cell lines. We define and test a new collection based on gene co-expression in cancer cell lines to compare the performance of the standard collections to an externally derived cell line based collection. The results show that GSEA findings vary significantly depending on the collection chosen for analysis. Henceforth, collections should be carefully selected and reported in studies that leverage GSEA.

Background: Epithelial-stromal crosstalk plays a critical role in invasive breast cancer pathogenesis; however, little is known on a systems level about how epithelial-stromal interactions evolve during carcinogenesis. Results: We develop a framework for building genome-wide epithelial-stromal co-expression networks composed of pairwise co-expression relationships between mRNA levels of genes expressed in the epithelium and stroma across a population of patients. We apply this method to laser capture micro-dissection expression profiling datasets in the setting of breast carcinogenesis. Our analysis shows that epithelial-stromal co-expression networks undergo extensive rewiring during carcinogenesis, with the emergence of distinct network hubs in normal breast, and estrogen receptor-positive and estrogen receptor-negative invasive breast cancer, and the emergence of distinct patterns of functional network enrichment. In contrast to normal breast, the strongest epithelial-stromal co-expression relationships in invasive breast cancer mostly represent self-loops, in which the same gene is co-expressed in epithelial and stromal regions. We validate this observation using an independent laser capture micro-dissection dataset and confirm that self-loop interactions are significantly increased in cancer by performing computational image analysis of epithelial and stromal protein expression using images from the Human Protein Atlas. Conclusions: Epithelial-stromal co-expression network analysis represents a new approach for systems-level analyses of spatially localized transcriptomic data. The analysis provides new biological insights into the rewiring of epithelial-stromal co-expression networks and the emergence of epithelial-stromal co-expression self-loops in breast cancer. The approach may facilitate the development of new diagnostics and therapeutics targeting epithelial-stromal interactions in cancer. Electronic supplementary material The online version of this article (doi:10.1186/s13059-015-0675-4) contains supplementary material, which is available to authorized users.

Introduction: Estrogen receptor (ER) and progesterone receptor (PR) testing are performed in the evaluation of breast cancer. While the clinical utility of ER as a predictive biomarker to identify patients likely to benefit from hormonal therapy is well-established, the added value of PR is less well-defined. The primary goals of our study were to assess the distribution, inter-assay reproducibility, and prognostic significance of breast cancer subtypes defined by patterns of ER and PR expression. Methods: We integrated gene expression microarray (GEM) and clinico-pathologic data from 20 published studies to determine the frequency (n = 4,111) and inter-assay reproducibility (n = 1,752) of ER/PR subtypes (ER+/PR+, ER+/PR-, ER-/PR-, ER-/PR+). To extend our findings, we utilized a cohort of patients from the Nurses’ Health Study (NHS) with ER/PR data recorded in the medical record and assessed on tissue microarrays (n = 2,011). In both datasets, we assessed the association of ER and PR expression with survival. Results: In a genome-wide analysis, progesterone receptor was among the least variable genes in ER- breast cancer. The ER-/PR+ subtype was rare (approximately 1 to 4%) and showed no significant reproducibility (Kappa = 0.02 and 0.06, in the GEM and NHS datasets, respectively). The vast majority of patients classified as ER-/PR+ in the medical record (97% and 94%, in the GEM and NHS datasets) were re-classified by a second method. In the GEM dataset (n = 2,731), progesterone receptor mRNA expression was associated with prognosis in ER+ breast cancer (adjusted P <0.001), but not in ER- breast cancer (adjusted P = 0.21). PR protein expression did not contribute significant prognostic information to multivariate models considering ER and other standard clinico-pathologic features in the GEM or NHS datasets. Conclusion: ER-/PR+ breast cancer is not a reproducible subtype. PR expression is not associated with prognosis in ER- breast cancer, and PR does not contribute significant independent prognostic information to multivariate models considering ER and other standard clinico-pathologic factors. Given that PR provides no clinically actionable information in ER+ breast cancer, these findings question the utility of routine PR testing in breast cancer.

A major goal in translational cancer research is to identify biological signatures driving cancer progression and metastasis. A common technique applied in genomics research is to cluster patients using gene expression data from a candidate prognostic gene set, and if the resulting clusters show statistically significant outcome stratification, to associate the gene set with prognosis, suggesting its biological and clinical importance. Recent work has questioned the validity of this approach by showing in several breast cancer data sets that ‘‘random’’ gene sets tend to cluster patients into prognostically variable subgroups. This work suggests that new rigorous statistical methods are needed to identify biologically informative prognostic gene sets. To address this problem, we developed Significance Analysis of Prognostic Signatures (SAPS) which integrates standard prognostic tests with a new prognostic significance test based on stratifying patients into prognostic subtypes with random gene sets. SAPS ensures that a significant gene set is not only able to stratify patients into prognostically variable groups, but is also enriched for genes showing strong univariate associations with patient prognosis, and performs significantly better than random gene sets. We use SAPS to perform a large meta-analysis (the largest completed to date) of prognostic pathways in breast and ovarian cancer and their molecular subtypes. Our analyses show that only a small subset of the gene sets found statistically significant using standard measures achieve significance by SAPS. We identify new prognostic signatures in breast and ovarian cancer and their corresponding molecular subtypes, and we show that prognostic signatures in ER negative breast cancer are more similar to prognostic signatures in ovarian cancer than to prognostic signatures in ER positive breast cancer. SAPS is a powerful new method for deriving robust prognostic biological signatures from clinically annotated genomic datasets.

In 2013, we published a comparative analysis of mutation and gene expression profiles and drug sensitivity measurements for 15 drugs characterized in the 471 cancer cell lines screened in the Genomics of Drug Sensitivity in Cancer (GDSC) and Cancer Cell Line Encyclopedia (CCLE). While we found good concordance in gene expression profiles, there was substantial inconsistency in the drug responses reported by the GDSC and CCLE projects. We received extensive feedback on the comparisons that we performed. This feedback, along with the release of new data, prompted us to revisit our initial analysis. We present a new analysis using these expanded data, where we address the most significant suggestions for improvements on our published analysis — that targeted therapies and broad cytotoxic drugs should have been treated differently in assessing consistency, that consistency of both molecular profiles and drug sensitivity measurements should be compared across cell lines, and that the software analysis tools provided should have been easier to run, particularly as the GDSC and CCLE released additional data. Our re-analysis supports our previous finding that gene expression data are significantly more consistent than drug sensitivity measurements. Using new statistics to assess data consistency allowed identification of two broad effect drugs and three targeted drugs with moderate to good consistency in drug sensitivity data between GDSC and CCLE. For three other targeted drugs, there were not enough sensitive cell lines to assess the consistency of the pharmacological profiles. We found evidence of inconsistencies in pharmacological phenotypes for the remaining eight drugs. Overall, our findings suggest that the drug sensitivity data in GDSC and CCLE continue to present challenges for robust biomarker discovery. This re-analysis provides additional support for the argument that experimental standardization and validation of pharmacogenomic response will be necessary to advance the broad use of large pharmacogenomic screens.