Yongseok Park

MethylSig: a whole genome DNA methylation analysis pipeline

MOTIVATION DNA methylation plays critical roles in gene regulation and cellular specification without altering DNA sequences. The wide application of reduced representation bisulfite sequencing (RRBS) and whole genome bisulfite sequencing (bis-seq) opens the door to study DNA methylation at single CpG site resolution. One challenging question is how best to test for significant methylation differences between groups of biological samples in order to minimize false positive findings. RESULTS We present a statistical analysis package, methylSig, to analyse genome-wide methylation differences between samples from different treatments or disease groups. MethylSig takes into account both read coverage and biological variation by utilizing a beta-binomial approach across biological samples for a CpG site or region, and identifies relevant differences in CpG methylation. It can also incorporate local information to improve group methylation level and/or variance estimation for experiments with small sample size. A permutation study based on data from enhanced RRBS samples shows that methylSig maintains a well-calibrated type-I error when the number of samples is three or more per group. Our simulations show that methylSig has higher sensitivity compared with several alternative methods. The use of methylSig is illustrated with a comparison of different subtypes of acute leukemia and normal bone marrow samples. AVAILABILITY methylSig is available as an R package at http://sartorlab.ccmb.med.umich.edu/software. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.

Real-time individual predictions of prostate cancer recurrence using joint models

Patients who were previously treated for prostate cancer with radiation therapy are monitored at regular intervals using a laboratory test called Prostate Specific Antigen (PSA). If the value of the PSA test starts to rise, this is an indication that the prostate cancer is more likely to recur, and the patient may wish to initiate new treatments. Such patients could be helped in making medical decisions by an accurate estimate of the probability of recurrence of the cancer in the next few years. In this article, we describe the methodology for giving the probability of recurrence for a new patient, as implemented on a web-based calculator. The methods use a joint longitudinal survival model. The model is developed on a training dataset of 2386 patients and tested on a dataset of 846 patients. Bayesian estimation methods are used with one Markov chain Monte Carlo (MCMC) algorithm developed for estimation of the parameters from the training dataset and a second quick MCMC developed for prediction of the risk of recurrence that uses the longitudinal PSA measures from a new patient.

Proteomic Characterization of Head and Neck Cancer Patient–Derived Xenografts

Despite advances in treatment approaches for head and neck squamous cell carcinoma (HNSCC), survival rates have remained stagnant due to the paucity of preclinical models that accurately reflect the human tumor. Patient-derived xenografts (PDX) are an emerging model system where patient tumors are implanted directly into mice. Increased understanding of the application and limitations of PDXs will facilitate their rational use. Studies to date have not reported protein profiles of PDXs. Therefore, we developed a large cohort of HNSCC PDXs and found that tumor take rate was not influenced by the clinical, pathologic, or processing features. Protein expression profiles, from a subset of the PDXs, were characterized by reverse-phase protein array and the data was compared with The Cancer Genome Atlas HNSCC data. Cluster analysis revealed that HNSCC PDXs were more similar to primary HNSCC than to any other tumor type. Interestingly, while a significant fraction of proteins were expressed similarly in both primary HNSCC and PDXs, a subset of proteins/phosphoproteins were expressed at higher (or lower) levels in PDXs compared with primary HNSCC. These findings indicate that the proteome is generally conserved in PDXs, but mechanisms for both positive and negative model selection and/or differences in the stromal components exist. Implications: Proteomic characterization of HNSCC PDXs demonstrates potential drivers for model selection and provides a framework for improved utilization of this expanding model system. Mol Cancer Res; 14(3); 278–86. ©2015 AACR.

Genetic and Epigenetic Alterations of TERT Are Associated with Inferior Outcome in Adolescent and Young Adult Patients with Melanoma

Progression of melanoma to distant sites in adolescents and young adults (AYAs) is not reliably predicted by clinicopathologic criteria. TERT promoter mutations when combined with BRAF/NRAS mutations correlate with adverse outcome in adult melanoma. To determine the prognostic value of TERT alterations in AYA melanoma, we investigated the association of TERT promoter mutations, as well as promoter methylation, an epigenetic alteration also linked to TERT upregulation, with TERT mRNA expression and outcome using a well-characterized cohort of 27 patients with melanoma (ages 8–25, mean 20). TERT mRNA expression levels were significantly higher in tumors harboring TERT promoter mutation and/or hypermethylation than those without either aberration (P = 0.046). TERT promoter mutations alone did not predict adverse outcomes (P = 0.50), but the presence of TERT promoter methylation, alone or concurrent with promoter mutations, correlated with reduced recurrence-free survival (P = 0.001). These data suggest that genetic and epigenetic alterations of TERT are associated with TERT upregulation and may predict clinical outcomes in AYA melanoma. A more exhaustive understanding of the different molecular mechanisms leading to increased TERT expression may guide development of prognostic assays to stratify AYA melanoma patients according to clinical risk.

Active estrogen receptor-alpha signaling in ovarian cancer models and clinical specimens

Purpose: High-grade serous ovarian cancer (HGSOC) is an aggressive disease with few available targeted therapies. Despite high expression of estrogen receptor-alpha (ERα) in approximately 80% of HGSOC and some small but promising clinical trials of endocrine therapy, ERα has been understudied as a target in this disease. We sought to identify hormone-responsive, ERα-dependent HGSOC. Experimental Design: We characterized endocrine response in HGSOC cells across culture conditions [ two-dimensional (2D), three-dimensional (3D), forced suspension] and in patient-derived xenograft (PDX) explants, assessing proliferation and gene expression. Estrogen-regulated transcriptome data were overlapped with public datasets to develop a comprehensive panel of ERα target genes. Expression of this panel and ERα H-score were assessed in HGSOC samples from patients who received endocrine therapy. Time on endocrine therapy was used as a surrogate for clinical response. Results: Proliferation is ERα-regulated in HGSOC cells in vitro and in vivo, and is partly dependent on 3D context. Transcriptomic studies identified genes shared by cell lines and PDX explants as ERα targets. The selective ERα downregulator (SERD) fulvestrant is more effective than tamoxifen in blocking ERα action. ERα H-score is predictive of efficacy of endocrine therapy, and this prediction is further improved by inclusion of target gene expression, particularly IGFBP3. Conclusions: Laboratory models corroborate intertumor heterogeneity of endocrine response in HGSOC but identify features associated with functional ERα and endocrine responsiveness. Assessing ERα function (e.g., IGFBP3 expression) in conjunction with H-score may help select patients who would benefit from endocrine therapy. Preclinical data suggest that SERDs might be more effective than tamoxifen. Clin Cancer Res; 23(14); 3802–12. ©2017 AACR.

Statistical Challenges in Analyzing Methylation and Long-Range Chromosomal Interaction Data

With the rapid development of high-throughput technologies such as array and next generation sequencing, genome-wide, nucleotide-resolution epigenomic data are increasingly available. In recent years, there has been particular interest in data on DNA methylation and 3-dimensional (3D) chromosomal organization, which are believed to hold keys to understand biological mechanisms, such as transcription regulation, that are closely linked to human health and diseases. However, small sample size, complicated correlation structure, substantial noise, biases, and uncertainties, all present difficulties for performing statistical inference. In this review, we present an overview of the new technologies that are frequently utilized in studying DNA methylation and 3D chromosomal organization. We focus on reviewing recent developments in statistical methodologies designed for better interrogating epigenomic data, pointing out statistical challenges facing the field whenever appropriate.

Acute Pancreatitis Has a Long-term Deleterious Effect on Physical Health Related Quality of Life

BACKGROUND & AIMS: It is not clear how acute pancreatitis (AP) affects health related quality of life (HRQOL). We aimed to determine the long‐term independent effect of AP on physical and mental HRQOL. METHODS: We analyzed data from 91 patients (mean 52 years of age, 54% women) admitted with AP to the University of Pittsburgh Medical Center from 2011 to 2015 who responded to telephone surveys at a median of 14 months after hospital discharge (interquartile range, 12–16 months). Individuals who did not answer the telephone survey were sent a questionnaire by regular mail. Patients answered questions from the 12‐Item Short‐Form Survey, and answers were used to calculate mental component summary (MCS) and physical component summary (PCS) scores with norm‐based scoring (normal ≥50). HRQOL for these subjects was compared with that of age‐ and sex‐matched individuals without pancreatitis (1:2) identified from the North American Pancreatitis Study. We controlled for other covariates using multivariable regression analysis. RESULTS: At follow‐up, individuals with AP had a significantly lower PCS score (46.2 ± 11.8) than did control subjects (51.1 ± 9.5; P < .01), but a similar MCS score. A 4‐point reduction of the PCS was attributed to AP after controlling for sociodemographic factors and medical comorbidities. The only pancreatitis‐related factor associated with low PCS score was multisystem organ failure. Presence of abdominal pain, analgesic use, disability, and current smoking at the time of follow‐up were also associated with lower PCS scores. Etiology of AP, disease severity (by Revised Atlanta classification), use of nutritional support, and performance of pancreatic interventions did not affect HRQOL at follow‐up. CONCLUSIONS: In a 14‐month follow‐up of patients hospitalized with AP, we found a meaningful, independent, and deleterious effect of AP in the physical HRQOL of these patients, compared to individuals without AP. Further research is needed to determine the duration of this impairment and to evaluate the effects of modifying risk factors.

Meta-analytic principal component analysis in integrative omics application

Motivation With the prevalent usage of microarray and massively parallel sequencing, numerous high‐throughput omics datasets have become available in the public domain. Integrating abundant information among omics datasets is critical to elucidate biological mechanisms. Due to the high‐dimensional nature of the data, methods such as principal component analysis (PCA) have been widely applied, aiming at effective dimension reduction and exploratory visualization. Results In this article, we combine multiple omics datasets of identical or similar biological hypothesis and introduce two variations of meta‐analytic framework of PCA, namely MetaPCA. Regularization is further incorporated to facilitate sparse feature selection in MetaPCA. We apply MetaPCA and sparse MetaPCA to simulations, three transcriptomic meta‐analysis studies in yeast cell cycle, prostate cancer, mouse metabolism and a TCGA pan‐cancer methylation study. The result shows improved accuracy, robustness and exploratory visualization of the proposed framework. Availability and implementation An R package MetaPCA is available online. (http://tsenglab.biostat.pitt.edu/software.htm). Supplementary information Supplementary data are available at Bioinformatics online.

Integrating DNA methylation and hydroxymethylation data with the mint pipeline

DNA methylation (5mC) plays important roles in mammalian development, oncogenesis, treatment response, and responses to the environment. DNA hydroxymethylation (5hmC) is also an informative epigenetic mark with distinct roles in regulation and cancer. Gold-standard, widely used technologies (bisulfite conversion, followed by deep sequencing) cannot distinguish between 5mC and 5hmC. Therefore, additional experiments are required to differentiate the two marks, and in silico methods are needed to analyze, integrate, and interpret these data. We developed the Methylation INTegration (mint) pipeline to support the comprehensive analysis of bisulfite conversion and immunoprecipitation-based methylation and hydroxymethylation assays, with additional steps toward integration, visualization, and interpretation. The pipeline is available as both a command line and a Galaxy graphical user interface tool. Both implementations require minimal configuration while remaining flexible to experiment specific needs. Cancer Res; 77(21); e27-30. ©2017 AACR.

MetaOmics: analysis pipeline and browser-based software suite for transcriptomic meta-analysis

SUMMARY The rapid advances of omics technologies have generated abundant genomic data in public repositories and effective analytical approaches are critical to fully decipher biological knowledge inside these data. Meta-analysis combines multiple studies of a related hypothesis to improve statistical power, accuracy and reproducibility beyond individual study analysis. To date, many transcriptomic meta-analysis methods have been developed, yet few thoughtful guidelines exist. Here, we introduce a comprehensive analytical pipeline and browser-based software suite, called MetaOmics, to meta-analyze multiple transcriptomic studies for various biological purposes, including quality control, differential expression analysis, pathway enrichment analysis, differential co-expression network analysis, prediction, clustering and dimension reduction. The pipeline includes many public as well as >10 in-house transcriptomic meta-analytic methods with data-driven and biological-aim-driven strategies, hands-on protocols, an intuitive user interface and step-by-step instructions. AVAILABILITY AND IMPLEMENTATION MetaOmics is freely available at https://github.com/metaOmics/metaOmics. SUPPLEMENTARY INFORMATION Supplementary data are available at Bioinformatics online.