Ted Laderas

The consensus molecular subtypes of colorectal cancer

Colorectal cancer (CRC) is a frequently lethal disease with heterogeneous outcomes and drug responses. To resolve inconsistencies among the reported gene expression–based CRC classifications and facilitate clinical translation, we formed an international consortium dedicated to large-scale data sharing and analytics across expert groups. We show marked interconnectivity between six independent classification systems coalescing into four consensus molecular subtypes (CMSs) with distinguishing features: CMS1 (microsatellite instability immune, 14%), hypermutated, microsatellite unstable and strong immune activation; CMS2 (canonical, 37%), epithelial, marked WNT and MYC signaling activation; CMS3 (metabolic, 13%), epithelial and evident metabolic dysregulation; and CMS4 (mesenchymal, 23%), prominent transforming growth factor–β activation, stromal invasion and angiogenesis. Samples with mixed features (13%) possibly represent a transition phenotype or intratumoral heterogeneity. We consider the CMS groups the most robust classification system currently available for CRC—with clear biological interpretability—and the basis for future clinical stratification and subtype-based targeted interventions.

Metastability of a supercompressed fluid monolayer.

Previous studies showed that monomolecular films of extracted calf surfactant collapse at the equilibrium spreading pressure during quasi-static compressions but become metastable at much higher surface pressures when compressed faster than a threshold rate. To determine the mechanism by which the films become metastable, we studied single-component films of 1-palmitoyl-2-oleoyl phosphatidylcholine (POPC). Initial experiments confirmed similar metastability of POPC if compressed above a threshold rate. Measurements at different surface pressures then showed that rates of collapse, although initially increasing above the equilibrium spreading pressure, reached a sharply defined maximum and then slowed considerably. When heated, rapidly compressed films recovered their ability to collapse with no discontinuous change in area, arguing that the metastability does not reflect transition of the POPC film to a new phase. These observations indicate that in several respects, the supercompression of POPC monolayers resembles the supercooling of three-dimensional liquids toward a glass transition.

High throughput sequencing in mice: a platform comparison identifies a preponderance of cryptic SNPs

BackgroundAllelic variation is the cornerstone of genetically determined differences in gene expression, gene product structure, physiology, and behavior. However, allelic variation, particularly cryptic (unknown or not annotated) variation, is problematic for follow up analyses. Polymorphisms result in a high incidence of false positive and false negative results in hybridization based analyses and hinder the identification of the true variation underlying genetically determined differences in physiology and behavior. Given the proliferation of mouse genetic models (e.g., knockout models, selectively bred lines, heterogeneous stocks derived from standard inbred strains and wild mice) and the wealth of gene expression microarray and phenotypic studies using genetic models, the impact of naturally-occurring polymorphisms on these data is critical. With the advent of next-generation, high-throughput sequencing, we are now in a position to determine to what extent polymorphisms are currently cryptic in such models and their impact on downstream analyses.ResultsWe sequenced the two most commonly used inbred mouse strains, DBA/2J and C57BL/6J, across a region of chromosome 1 (171.6 – 174.6 megabases) using two next generation high-throughput sequencing platforms: Applied Biosystems (SOLiD) and Illumina (Genome Analyzer). Using the same templates on both platforms, we compared realignments and single nucleotide polymorphism (SNP) detection with an 80 fold average read depth across platforms and samples. While public datasets currently annotate 4,527 SNPs between the two strains in this interval, thorough high-throughput sequencing identified a total of 11,824 SNPs in the interval, including 7,663 new SNPs. Furthermore, we confirmed 40 missense SNPs and discovered 36 new missense SNPs.ConclusionComparisons utilizing even two of the best characterized mouse genetic models, DBA/2J and C57BL/6J, indicate that more than half of naturally-occurring SNPs remain cryptic. The magnitude of this problem is compounded when using more divergent or poorly annotated genetic models. This warrants full genomic sequencing of the mouse strains used as genetic models.

TandTRAQ: An open-source tool for integrated protein identification and quantitation

UNLABELLED Integrating qualitative protein identification with quantitative protein analysis is non-trivial, given incompatibility in output formats. We present TandTRAQ, a standalone utility that integrates results from i-Tracker, an open-source iTRAQ quantitation program with the search results from X?Tandem, an open-source proteome search engine. The utility runs from the command-line and can be easily integrated into a pipeline for automation. AVAILABILITY The TandTRAQ Perl scripts are freely available for download at http://www.ohsucancer.com/isrdev/tandtraq/

Computational detection of alternative exon usage

Background: With the advent of the GeneChip Exon Arrays, it is now possible to extract “exon-level” expression estimates, allowing for detection of alternative splicing events, one of the primary mechanisms of transcript diversity. In the context of (1) a complex trait use case and (2) a human cerebellum vs. heart comparison on previously validated data, we present a transcript-based statistical model and validation framework to allow detection of alternative exon usage (AEU) between different groups. To illustrate the approach, we detect and confirm differences in exon usage in the two of the most widely studied mouse genetic models (the C57BL/6J and DBA/2J inbred strains) and in a human dataset. Results: We developed a computational framework that consists of probe level annotation mapping and statistical modeling to detect putative AEU events, as well as visualization and alignment with known splice events. We show a dramatic improvement (∼25 fold) in the ability to detect these events using the appropriate annotation and statistical model which is actually specified at the transcript level, as compared with the transcript cluster/gene-level annotation used on the array. An additional component of this workflow is a probe index that allows ranking AEU candidates for validation and can aid in identification of false positives due to single nucleotide polymorphisms. Discussion: Our work highlights the importance of concordance between the functional unit interrogated (e.g., gene, transcripts) and the entity (e.g., exon, probeset) within the statistical model. The framework we present is broadly applicable to other platforms (including RNAseq).

Teaching data science fundamentals through realistic synthetic clinical cardiovascular data

Objective Our goal was to create a synthetic dataset and curricular materials to assist in teaching fundamentals of translational data science. Materials and Methods A literature review was conducted to extract current cardiovascular risk score logic, data elements, and population characteristics. Then, clinical data elements in the models were pulled from clinical data and transformed to the Observational Medical Outcomes Partnership (OMOP) common data model; genetic data elements were added based on population rates. A hybrid Bayesian network was used to create synthetic data from the logical elements of the risk scores and the underlying population frequencies of the clinical data. Results A synthetic dataset of 446,000 patients was created. A two-day curriculum was created based on this synthetic data with exploratory data analysis and machine learning components. The curriculum was offered on two separate occasions; the two groups of learners were given the curriculum and data, and results were tallied, summarized, and compared. Students’ ability to complete the challenge was mixed; more experienced students achieved a range of 70%-85% in balanced accuracy, but many others did not perform better than the baseline model. Discussion Overall, students enjoyed the course and dataset, but some struggled to consistently apply machine learning techniques. The curriculum, data set, techniques for generation, and results are available for others to use for their own training. Conclusion A realistic synthetic data with clinical and genetic components helps students learn issues in cardiovascular risk scoring, practice data science skills, and compete in a challenge to improve identification of risk.

Integrated functional and mass spectrometry-based flow cytometric phenotyping to describe the immune microenvironment in acute myeloid leukemia

A hallmark of the development of cancer is its ability to avoid detection and elimination by the immune system. There are many identified mechanisms of this immune evasion that can be measured both phenotypically and functionally. Functional studies directly show the ability of the tumor microenvironment to suppress immune responses, typically measured as lymphocyte proliferation, cytokine production or killing ability. While a direct measurement of function is ideal, these assays require ex vivo activation which may not accurately mimic in vivo conditions. Phenotypic assays can directly measure the distribution and activation of immune cell types rapidly after isolation, preserving the conditions present in the patient. While conventional flow cytometry is a rapid and well established assay, it currently allows for measurement of only 12-14 parameters. Mass spectrometry-based flow cytometry, or CyTOF, offers the ability to measure 3-fold more parameters than conventional optical-based modalities providing an advantage in depth of analysis that can be crucial for precious human samples. The goal of this report is to describe the system our group has developed to measure both the phenotype and function of immune cells in the bone marrow of patients with acute myeloid leukemia. We hope to explain our system in the context of previous studies aimed at measuring immune status in tumors and to inform the reader as to some experimental approaches our group has found useful in developing the basic data required to rationally pursue immune-based therapies for patients with cancer.

Between Pathways and Networks Lies Context: Implications for Precision Medicine

Precision medicine, broadly defined as considering individual variability in genes, environment, and lifestyle for each person in disease prevention and selection of suitable medical intervention, shows strong promise in the treatment of cancer. Selecting therapies is complicated by multiple routes to gene dysregulation, which manifest in the individual patient within the many different types of genomic measurements. Additionally, multiple mutations exist in patients, a phenomenon known as oncogenic collaboration, which further complicates the selection of therapy. In this article, we discuss current approaches using biological pathways and networks to unify the many types of OMICs data. We argue that a contextual approach combining cancer pathways and networks could lead to a proper understanding of the biology of this significant disease.

A Network-Based Model of Oncogenic Collaboration for Prediction of Drug Sensitivity

Tumorigenesis is a multi-step process, involving the acquisition of multiple oncogenic mutations that transform cells, resulting in systemic dysregulation that enables proliferation, invasion, and other cancer hallmarks. The goal of precision medicine is to identify therapeutically-actionable mutations from large-scale omic datasets. However, the multiplicity of oncogenes required for transformation, known as oncogenic collaboration, makes assigning effective treatments difficult. Motivated by this observation, we propose a new type of oncogenic collaboration where mutations in genes that interact with an oncogene may contribute to the oncogene’s deleterious potential, a new genomic feature that we term “surrogate oncogenes.” Surrogate oncogenes are representatives of these mutated subnetworks that interact with oncogenes. By mapping mutations to a protein–protein interaction network, we determine the significance of the observed distribution using permutation-based methods. For a panel of 38 breast cancer cell lines, we identified a significant number of surrogate oncogenes in known oncogenes such as BRCA1 and ESR1, lending credence to this approach. In addition, using Random Forest Classifiers, we show that these significant surrogate oncogenes predict drug sensitivity for 74 drugs in the breast cancer cell lines with a mean error rate of 30.9%. Additionally, we show that surrogate oncogenes are predictive of survival in patients. The surrogate oncogene framework incorporates unique or rare mutations from a single sample, and therefore has the potential to integrate patient-unique mutations into drug sensitivity predictions, suggesting a new direction in precision medicine and drug development. Additionally, we show the prevalence of significant surrogate oncogenes in multiple cancers from The Cancer Genome Atlas, suggesting that surrogate oncogenes may be a useful genomic feature for guiding pancancer analyses and assigning therapies across many tissue types.

Abstract 603: Consensus molecular subtyping through a community of experts advances unsupervised gene expression-based disease classification and facilitates clinical translation

Background: Gene expression-based subtyping is widely accepted as a relevant source of disease stratification. Despite the widespread use, its translational and clinical utility is hampered by discrepant results, likely related to differences in data processing and algorithms applied to diverse patient cohorts, sample preparation methods, and gene expression platforms. In the absence of a clear methodological gold standard to perform such analyses, a more general framework that integrates and compares multiple strategies is needed to define common disease patterns in a principled, unbiased manner. Methods: We formed a consortium of 6 independent experts groups - each with a previously published CRC classifier, ranging from 3 to 6 subtypes - to understand similarities and differences of their subtyping systems. Sage Bionetworks functioned as neutral party to aggregate public and proprietary data (Synapse platform) and perform meta-analysis. Each group applied its CRC subtyping signature to the collection of data sets with gene expression (n = 4,151, predominantly stage II and III). Using the resulting subtype labels, we developed a network-based model and applied a Markov cluster algorithm to detect robust network substructures that would indicate recurring subtype patterns and therefore a consensus subtyping system. Correlative analyses using clinico-pathological, genomic and epigenomic features was performed to robustly characterize the identified subtypes. Results: This analytical framework revealed significant interconnectivity between the six independent classification systems, leading to the identification of four biologically distinct consensus molecular subtypes (CMS) enriched for key pathway traits: CMS1 (MSI Immune), hypermutated, microsatellite unstable, with strong immune activation; CMS2 (Canonical), epithelial, chromosomally unstable, with marked WNT and MYC signaling activation; CMS3 (Metabolic), epithelial, with evident metabolic dysregulation; and CMS4 (Mesenchymal), prominent TGFβ activation, angiogenesis, stromal invasion. Patients diagnosed with MSI Immune tumors had worse survival after relapse and those with mesenchymal tumors had increased risk of metastasis and worse overall survival. Discussion: We describe a novel methodological paradigm for deriving benchmarks of disease subtyping. Our work represents the first example of a community of experts identifying and advocating for a single reproducible model for cancer subtyping, effectively unifying previous classifiers. In the CRC domain, the uniformity afforded by this new classification system and its application to a large data set revealed important subtype-specific biological associations that were previously unnoticed or marginally significant, supporting a new taxonomy of the disease. Citation Format: Justin Guinney, Rodrigo Dienstmann, Xin Wang, Aurelien de Reynies, Andreas Schlicker, Charlotte Soneson, Laetitia Marisa, Paul Roepman, Gift Nyamundanda, Paolo Angelino, Brian Bot, Jeffrey S. Morris, Iris Simon, Sarah Gerster, Evelyn Fessler, Felipe de Sousa e Melo, Edoardo Missiaglia, Hena Ramay, David Barras, Krisztian Homicsko, Dipen Maru, Ganiraju Manyam, Bradley Broom, Valerie Boige, Ted Laderas, Ramon Salazar, Joe W. Gray, Josep Tabernero, Rene Bernards, Stephen Friend, Pierre Laurent-Puig, Jan P. Medema, Anguraj Sadanandam, Lodewyk Wessels, Mauro Delorenzi, Scott Kopetz, Louis Vermeulen, Sabine Tejpar. Consensus molecular subtyping through a community of experts advances unsupervised gene expression-based disease classification and facilitates clinical translation. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 603. doi:10.1158/1538-7445.AM2015-603