Andrew E. Jaffe

The sva package for removing batch effects and other unwanted variation in high-throughput experiments

Heterogeneity and latent variables are now widely recognized as major sources of bias and variability in high-throughput experiments. The most well-known source of latent variation in genomic experiments are batch effects-when samples are processed on different days, in different groups or by different people. However, there are also a large number of other variables that may have a major impact on high-throughput measurements. Here we describe the sva package for identifying, estimating and removing unwanted sources of variation in high-throughput experiments. The sva package supports surrogate variable estimation with the sva function, direct adjustment for known batch effects with the ComBat function and adjustment for batch and latent variables in prediction problems with the fsva function.

Accounting for cellular heterogeneity is critical in epigenome-wide association studies

BackgroundEpigenome-wide association studies of human disease and other quantitative traits are becoming increasingly common. A series of papers reporting age-related changes in DNA methylation profiles in peripheral blood have already been published. However, blood is a heterogeneous collection of different cell types, each with a very different DNA methylation profile.ResultsUsing a statistical method that permits estimating the relative proportion of cell types from DNA methylation profiles, we examine data from five previously published studies, and find strong evidence of cell composition change across age in blood. We also demonstrate that, in these studies, cellular composition explains much of the observed variability in DNA methylation. Furthermore, we find high levels of confounding between age-related variability and cellular composition at the CpG level.ConclusionsOur findings underscore the importance of considering cell composition variability in epigenetic studies based on whole blood and other heterogeneous tissue sources. We also provide software for estimating and exploring this composition confounding for the Illumina 450k microarray.

DNA methylation age of blood predicts all-cause mortality in later life

BackgroundDNA methylation levels change with age. Recent studies have identified biomarkers of chronological age based on DNA methylation levels. It is not yet known whether DNA methylation age captures aspects of biological age.ResultsHere we test whether differences between people’s chronological ages and estimated ages, DNA methylation age, predict all-cause mortality in later life. The difference between DNA methylation age and chronological age (Δage) was calculated in four longitudinal cohorts of older people. Meta-analysis of proportional hazards models from the four cohorts was used to determine the association between Δage and mortality. A 5-year higher Δage is associated with a 21% higher mortality risk, adjusting for age and sex. After further adjustments for childhood IQ, education, social class, hypertension, diabetes, cardiovascular disease, and APOE e4 status, there is a 16% increased mortality risk for those with a 5-year higher Δage. A pedigree-based heritability analysis of Δage was conducted in a separate cohort. The heritability of Δage was 0.43.ConclusionsDNA methylation-derived measures of accelerated aging are heritable traits that predict mortality independently of health status, lifestyle factors, and known genetic factors.

Multiple loci influence erythrocyte phenotypes in the CHARGE Consortium

Measurements of erythrocytes within the blood are important clinical traits and can indicate various hematological disorders. We report here genome-wide association studies (GWAS) for six erythrocyte traits, including hemoglobin concentration (Hb), hematocrit (Hct), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC) and red blood cell count (RBC). We performed an initial GWAS in cohorts of the CHARGE Consortium totaling 24,167 individuals of European ancestry and replication in additional independent cohorts of the HaemGen Consortium totaling 9,456 individuals. We identified 23 loci significantly associated with these traits in a meta-analysis of the discovery and replication cohorts (combined P values ranging from 5 × 10−8 to 7 × 10−86). Our findings include loci previously associated with these traits (HBS1L-MYB, HFE, TMPRSS6, TFR2, SPTA1) as well as new associations (EPO, TFRC, SH2B3 and 15 other loci). This study has identified new determinants of erythrocyte traits, offering insight into common variants underlying variation in erythrocyte measures.

Mapping DNA methylation across development, genotype and schizophrenia in the human frontal cortex

DNA methylation (DNAm) is important in brain development and is potentially important in schizophrenia. We characterized DNAm in prefrontal cortex from 335 non-psychiatric controls across the lifespan and 191 patients with schizophrenia and identified widespread changes in the transition from prenatal to postnatal life. These DNAm changes manifest in the transcriptome, correlate strongly with a shifting cellular landscape and overlap regions of genetic risk for schizophrenia. A quarter of published genome-wide association studies (GWAS)-suggestive loci (4,208 of 15,930, P < 10−100) manifest as significant methylation quantitative trait loci (meQTLs), including 59.6% of GWAS-positive schizophrenia loci. We identified 2,104 CpGs that differ between schizophrenia patients and controls that were enriched for genes related to development and neurodifferentiation. The schizophrenia-associated CpGs strongly correlate with changes related to the prenatal-postnatal transition and show slight enrichment for GWAS risk loci while not corresponding to CpGs differentiating adolescence from later adult life. These data implicate an epigenetic component to the developmental origins of this disorder.

Global DNA hypomethylation is associated with in utero exposure to cotinine and perfluorinated alkyl compounds.

Environmental exposures in-utero may alter the epigenome, thus impacting chromosomal stability and gene expression. We hypothesized that in utero exposures to maternal smoking and perfluoroalkyl compounds (PFCs) are associated with global DNA hypomethylation in umbilical cord serum. Our objective was to determine if global DNA methylation could be used as a biomarker of in utero exposures to maternal smoking and PFCs. Using an ELISA-based method, global DNA methylation was quantified in umbilical cord serum from 30 newborns with high (>10 ng/ml, mean 123.8 ng/ml), low (range 1-10 ng/ml, mean 1.6 ng/ml) and very low (<1 ng/ml, mean 0.06 ng/ml) cord serum cotinine levels. Y chromosome analysis was performed to rule out maternal DNA cross-contamination. Cord serum global DNA methylation showed an inverse dose response to serum cotinine levels (p<0.001). Global DNA methylation levels in cord blood were the lowest among newborns with smoking mothers (mean=15.04%; 95% CI, 8.4, 21.7) when compared to babies of mothers who were second-hand smokers (21.1%; 95% CI, 16.6, 25.5) and non-smokers (mean=29.2%; 95% CI, 20.1, 38.1). Global DNA methylation was inversely correlated with serum PFOA (r= -0.72, p <0.01) but not PFOS levels. Serum Y chromosome analyses did not detect maternal DNA cross-contamination. This study supports the use of global DNA methylation status as a biomarker of in utero exposure to cigarette smoke and PFCs.

Genome-Wide Analysis of Promoter Methylation Associated with Gene Expression Profile in Pancreatic Adenocarcinoma

Purpose: The goal of this study was to comprehensively identify CpG island methylation alterations between pancreatic cancers and normal pancreata and their associated gene expression alterations. Experimental Design: We employed methylated CpG island amplification followed by CpG island microarray, a method previously validated for its accuracy and reproducibility, to analyze the methylation profile of 27,800 CpG islands covering 21 MB of the human genome in nine pairs of pancreatic cancer versus normal pancreatic epithelial tissues and in three matched pairs of pancreatic cancer versus lymphoid tissues from the same individual. Results: This analysis identified 1,658 known loci that were commonly differentially methylated in pancreatic cancer compared with normal pancreas. By integrating the pancreatic DNA methylation status with the gene expression profiles of the same samples before and after treatment with the DNA methyltransferase inhibitor 5-aza-2′-deoxycytidine, and the histone deacetylase inhibitor, trichostatin A, we identified dozens of aberrantly methylated and differentially expressed genes in pancreatic cancers including a more comprehensive list of hypermethylated and silenced genes that have not been previously described as targets for aberrant methylation in cancer. Conclusion: We expected that the identification of aberrantly hypermethylated and silenced genes will have diagnostic, prognostic, and therapeutic applications. Clin Cancer Res; 17(13); 4341–54. ©2011 AACR.

The PsychENCODE project

Recent research on disparate psychiatric disorders has implicated rare variants in genes involved in global gene regulation and chromatin modification, as well as many common variants located primarily in regulatory regions of the genome. Understanding precisely how these variants contribute to disease will require a deeper appreciation for the mechanisms of gene regulation in the developing and adult human brain. The PsychENCODE project aims to produce a public resource of multidimensional genomic data using tissue- and cell type–specific samples from approximately 1,000 phenotypically well-characterized, high-quality healthy and disease-affected human post-mortem brains, as well as functionally characterize disease-associated regulatory elements and variants in model systems. We are beginning with a focus on autism spectrum disorder, bipolar disorder and schizophrenia, and expect that this knowledge will apply to a wide variety of psychiatric disorders. This paper outlines the motivation and design of PsychENCODE.

Developmental regulation of human cortex transcription and its clinical relevance at single base resolution

Transcriptome analysis of human brain provides fundamental insight into development and disease, but it largely relies on existing annotation. We sequenced transcriptomes of 72 prefrontal cortex samples across six life stages and identified 50,650 differentially expression regions (DERs) associated with developmental and aging, agnostic of annotation. While many DERs annotated to non-exonic sequence (41.1%), most were similarly regulated in cytosolic mRNA extracted from independent samples. The DERs were developmentally conserved across 16 brain regions and in the developing mouse cortex, and were expressed in diverse cell and tissue types. The DERs were further enriched for active chromatin marks and clinical risk for neurodevelopmental disorders such as schizophrenia. Lastly, we demonstrate quantitatively that these DERs associate with a changing neuronal phenotype related to differentiation and maturation. These data show conserved molecular signatures of transcriptional dynamics across brain development, have potential clinical relevance and highlight the incomplete annotation of the human brain transcriptome.

Ballgown bridges the gap between transcriptome assembly and expression analysis.

Analysis of raw reads from RNA sequencing (RNA-seq) makes it possible to reconstruct complete gene structures, including multiple splice variants, without relying on previously established annotations1–3. Downstream statistical modeling of summarized gene or transcript expression data output from these pipelines is facilitated by the Bioconductor project, which provides open-source tools for analysis of high-throughput genomics data4. However, the outputs of upstream processing tools often are aggregated across samples or are not in a format that is readily compatible with downstream Bioconductor packages. This gap has slowed rigorous statistical analysis of expression quantitative trait locus (eQTL), time-course, continuous covariates or of confounded experimental designs at the transcript level and has led to considerable controversy in the analysis of population-level RNA-seq data5. In this Correspondence, we report the development of two pieces of software, Tablemaker and Ballgown, that bridge the gap between transcriptome assembly and fast, flexible differential expression analysis (Supplementary Fig. 1).