Biomedical subjects

Steve Horvath

Publications and source records attributed to Steve Horvath.

At least 19 recordsLinked to original sources

PubMed2026

Epigenetic Clocks of Biological Aging and Cognitively Healthy Longevity: The Women's Health Initiative Memory Study.

BACKGROUND: Little is known about whether epigenetic age acceleration (EAA) clocks are capable of predicting exceptional longevity with or without preserved cognitive function. METHODS: We examined 5844 women from the Women's Health Initiative Memory Study. Fifteen epigenetic clocks were measured at baseline (1996-1999). Longevity outcomes were defined as: 1) survival to age 90 with preserved cognition (n = 1726, 29.5%); or 2) survival to age 90 with cognitive impairment (n = 956, 16.4%); vs. 3) death before age 90 (n = 2611, 44.7%). Logistic regression models examined associations between the 15 clocks and survival to age 90 (vs. death before age 90), adjusting for covariates. Multinomial logistic regression models examined associations with survival to age 90 without cognitive impairment and survival to age 90 with cognitive impairment (each vs. death before age 90), also adjusting for covariates. RESULTS: Each standard deviation increase in EAA for the first-generation clocks was associated with 7%-18% reduced odds of survival to age 90 vs. earlier death. Stronger associations were observed for second- and third-generation clocks, including AgeAccelGrim2 (OR = 0.66; 95% CI 0.61-0.71), PCGrimAge (OR = 0.64; 95% CI 0.59-0.69), PCPhenoAge (OR = 0.73; 95% CI 0.68-0.78) and DunedinPACE (OR = 0.77; 95% CI 0.72-0.82). None of the clocks was more strongly associated with survival to age 90 with preserved cognition than with survival to age 90 with cognitive impairment, relative to death before age 90. CONCLUSION: All epigenetic clocks were associated with exceptional longevity, but none were associated with cognitive healthspan. Developing clocks that can differentiate long survival with and without preserved cognitive function is critical.

Healthspan↗

PubMed2024

A methylation risk score for chronic kidney disease: a HyperGEN study.

Chronic kidney disease (CKD) impacts about 1 in 7 adults in the United States, but African Americans (AAs) carry a disproportionately higher burden of disease. Epigenetic modifications, such as DNA methylation at cytosine-phosphate-guanine (CpG) sites, have been linked to kidney function and may have clinical utility in predicting the risk of CKD. Given the dynamic relationship between the epigenome, environment, and disease, AAs may be especially sensitive to environment-driven methylation alterations. Moreover, risk models incorporating CpG methylation have been shown to predict disease across multiple racial groups. In this study, we developed a methylation risk score (MRS) for CKD in cohorts of AAs. We selected nine CpG sites that were previously reported to be associated with estimated glomerular filtration rate (eGFR) in epigenome-wide association studies to construct a MRS in the Hypertension Genetic Epidemiology Network (HyperGEN). In logistic mixed models, the MRS was significantly associated with prevalent CKD and was robust to multiple sensitivity analyses, including CKD risk factors. There was modest replication in validation cohorts. In summary, we demonstrated that an eGFR-based CpG score is an independent predictor of prevalent CKD, suggesting that MRS should be further investigated for clinical utility in evaluating CKD risk and progression.

Humans↗

PubMed2022

Significant variation in the performance of DNA methylation predictors across data preprocessing and normalization strategies.

BACKGROUND: DNA methylation (DNAm)-based predictors hold great promise to serve as clinical tools for health interventions and disease management. While these algorithms often have high prediction accuracy, the consistency of their performance remains to be determined. We therefore conduct a systematic evaluation across 101 different DNAm data preprocessing and normalization strategies and assess how each analytical strategy affects the consistency of 41 DNAm-based predictors. RESULTS: Our analyses are conducted in a large EPIC DNAm array dataset from the Jackson Heart Study (N = 2053) that included 146 pairs of technical replicate samples. By estimating the average absolute agreement between replicate pairs, we show that 32 out of 41 predictors (78%) demonstrate excellent consistency when appropriate data processing and normalization steps are implemented. Across all pairs of predictors, we find a moderate correlation in performance across analytical strategies (mean rho = 0.40, SD = 0.27), highlighting significant heterogeneity in performance across algorithms. Successful or unsuccessful removal of technical variation furthermore significantly impacts downstream phenotypic association analysis, such as all-cause mortality risk associations. CONCLUSIONS: We show that DNAm-based algorithms are sensitive to technical variation. The right choice of data processing strategy is important to achieve reproducible estimates and improve prediction accuracy in downstream phenotypic association analyses. For each of the 41 DNAm predictors, we report its degree of consistency and provide the best performing analytical strategy as a guideline for the research community. As DNAm-based predictors become more and more widely used, our work helps improve their performance and standardize their implementation.

DNA Methylation↗

PubMed2006

Network neighborhood analysis with the multi-node topological overlap measure.

MOTIVATION: The goal of neighborhood analysis is to find a set of genes (the neighborhood) that is similar to an initial 'seed' set of genes. Neighborhood analysis methods for network data are important in systems biology. If individual network connections are susceptible to noise, it can be advantageous to define neighborhoods on the basis of a robust interconnectedness measure, e.g. the topological overlap measure. Since the use of multiple nodes in the seed set may lead to more informative neighborhoods, it can be advantageous to define multi-node similarity measures. RESULTS: The pairwise topological overlap measure is generalized to multiple network nodes and subsequently used in a recursive neighborhood construction method. A local permutation scheme is used to determine the neighborhood size. Using four network applications and a simulated example, we provide empirical evidence that the resulting neighborhoods are biologically meaningful, e.g. we use neighborhood analysis to identify brain cancer related genes. AVAILABILITY: An executable Windows program and tutorial for multi-node topological overlap measure (MTOM) based analysis can be downloaded from the webpage (http://www.genetics.ucla.edu/labs/horvath/MTOM/).

Algorithms↗

PubMed2006

Conservation and evolution of gene coexpression networks in human and chimpanzee brains.

Comparisons of gene expression between human and non-human primate brains have identified hundreds of differentially expressed genes, yet translating these lists into key functional distinctions between species has proved difficult. Here we provide a more integrated view of human brain evolution by examining the large-scale organization of gene coexpression networks in human and chimpanzee brains. We identify modules of coexpressed genes that correspond to discrete brain regions and quantify their conservation between the species. Module conservation in cerebral cortex is significantly weaker than module conservation in subcortical brain regions, revealing a striking gradient that parallels known evolutionary hierarchies. We introduce a method for identifying species-specific network connections and demonstrate how differential network connectivity can be used to identify key drivers of evolutionary change. By integrating our results with comparative genomic sequence data and estimates of protein sequence divergence rates, we confirm a number of network predictions and validate these findings. Our results provide insights into the molecular bases of primate brain organization and demonstrate the general utility of weighted gene coexpression network analysis.

Animals↗

PubMed2006

Repetitive sequence environment distinguishes housekeeping genes.

Housekeeping genes are expressed across a wide variety of tissues. Since repetitive sequences have been reported to influence the expression of individual genes, we employed a novel approach to determine whether housekeeping genes can be distinguished from tissue-specific genes by their repetitive sequence context. We show that Alu elements are more highly concentrated around housekeeping genes while various longer (>400-bp) repetitive sequences ("repeats"), including Long Interspersed Nuclear Element-1 (LINE-1) elements, are excluded from these regions. We further show that isochore membership does not distinguish housekeeping genes from tissue-specific genes and that repetitive sequence environment distinguishes housekeeping genes from tissue-specific genes in every isochore. The distinct repetitive sequence environment, in combination with other previously published sequence properties of housekeeping genes, was used to develop a method of predicting housekeeping genes on the basis of DNA sequence alone. Using expression across tissue types as a measure of success, we demonstrate that repetitive sequence environment is by far the most important sequence feature identified to date for distinguishing housekeeping genes.

3' Flanking Region↗

PubMed2006

Identification of inflammatory gene modules based on variations of human endothelial cell responses to oxidized lipids.

Oxidized phospholipids are thought to promote atherogenesis by stimulating endothelial cells (ECs) to produce inflammatory cytokines, such as IL-8. In studies with mouse models, we previously demonstrated that genetic variation in inflammatory responses of endothelial cells to oxidized lipids contributes importantly to atherosclerosis susceptibility. We now show that similar variations occur in cultured aortic ECs derived from multiple heart transplant donors. These variations were stably maintained between passages and, thus, reflect either genetic or epigenetic regulatory differences. Expression array analysis of aortic EC cultures derived from 12 individuals revealed that >1,000 genes were regulated by oxidized phospholipids. We have used the observed variations in the sampled population to construct a gene coexpression network comprised of 15 modules of highly connected genes. We show that several identified modules are significantly enriched in genes for known pathways and confirm a module enriched for unfolded protein response (UPR) genes using siRNA and the UPR inducer tunicamycin. On the basis of the constructed network, we predicted that a gene of unknown function (MGC4504) present in the UPR module is a target for UPR transcriptional activator ATF4. Our data also indicate that IL-8 is present in the UPR module and is regulated, in part, by the UPR. We validate these by using siRNA. In conclusion, we show that interindividual variability can be used to group genes into pathways and predict gene-gene regulatory relationships, thus identifying targets potentially involved in susceptibility to common diseases such as atherosclerosis.

Activating Transcription Factor 4↗

PubMed2006

Integrating genetic and network analysis to characterize genes related to mouse weight.

Systems biology approaches that are based on the genetics of gene expression have been fruitful in identifying genetic regulatory loci related to complex traits. We use microarray and genetic marker data from an F2 mouse intercross to examine the large-scale organization of the gene co-expression network in liver, and annotate several gene modules in terms of 22 physiological traits. We identify chromosomal loci (referred to as module quantitative trait loci, mQTL) that perturb the modules and describe a novel approach that integrates network properties with genetic marker information to model gene/trait relationships. Specifically, using the mQTL and the intramodular connectivity of a body weight-related module, we describe which factors determine the relationship between gene expression profiles and weight. Our approach results in the identification of genetic targets that influence gene modules (pathways) that are related to the clinical phenotypes of interest.

Animals↗

PubMed2006

Gene connectivity, function, and sequence conservation: predictions from modular yeast co-expression networks.

BACKGROUND: Genes and proteins are organized into functional modular networks in which the network context of a gene or protein has implications for cellular function. Highly connected hub proteins, largely responsible for maintaining network connectivity, have been found to be much more likely to be essential for yeast survival. RESULTS: Here we investigate the properties of weighted gene co-expression networks formed from multiple microarray datasets. The constructed networks approximate scale-free topology, but this is not universal across all datasets. We show strong positive correlations between gene connectivity within the whole network and gene essentiality as well as gene sequence conservation. We demonstrate the preservation of a modular structure of the networks formed, and demonstrate that, within some of these modules, it is possible to observe a strong correlation between connectivity and essentiality or between connectivity and conservation within the modules particularly within modules containing larger numbers of essential genes. CONCLUSION: Application of these techniques can allow a finer scale prediction of relative gene importance for a particular process within a group of similarly expressed genes.

Base Sequence↗

PubMed2006

Flap endonuclease 1 is overexpressed in prostate cancer and is associated with a high Gleason score.

OBJECTIVE: To investigate the expression and potential clinical usefulness of structure-specific flap endonuclease 1 (FEN-1) in human primary prostate cancer using tissue microarray technology, as FEN-1 was recently identified to be overexpressed in CL1.1, the most aggressive clone generated from the hormone-refractory prostate cancer cell line CL1. MATERIALS AND METHODS: Immunohistochemistry was performed on tissue microarrays constructed from paraffin-embedded specimens of primary prostate cancer from 246 patients who had had a radical prostatectomy. Prostatic intraepithelial neoplasia (PIN), benign prostatic hyperplasia (BPH) and normal prostate epithelium were represented on the array. FEN-1 nuclear expression was scored based on the percentage of target cells staining positively, and correlated with Gleason score, preoperative prostate-specific antigen (PSA) level and pathological stage. The time to PSA recurrence was also analysed. RESULTS: The mean expression of FEN-1 was significantly higher in cancer (36.7%) than in normal (13.2%), BPH (4.5%) and PIN (15.4%) specimens (P < 0.001). FEN-1 expression was significantly correlated with Gleason score (ó = 0.23, P = 0.002). A higher preoperative serum PSA level (P = 0.015), Gleason score > or = 7 (P < 0.001), seminal vesicle invasion (P < 0.001) and capsular involvement (P = 0.004) were associated with PSA recurrence, whereas FEN-1 expression was not. In a multivariate analysis, only Gleason score > or = 7 (P < 0.001), seminal vesicle invasion (P = 0.005) and capsular involvement (P = 0.009) were retained as independent predictors for PSA recurrence. CONCLUSIONS: FEN-1 is overexpressed in prostate cancer compared with matched normal prostate, and its expression increases with tumour dedifferentiation, as shown by increasing Gleason score. These results suggest that FEN-1 might be a potential marker for selecting patients at high risk, and a potential target for prostate cancer diagnosis and therapy.

Aged↗

PubMed2005

Extreme skewing of X chromosome inactivation in mothers of homosexual men.

Human sexual preference is a sexually dimorphic trait with a substantial genetic component. Linkage of male sexual orientation to markers on the X chromosome has been reported in some families. Here, we measured X chromosome inactivation ratios in 97 mothers of homosexual men and 103 age-matched control women without gay sons. The number of women with extreme skewing of X-inactivation was significantly higher in mothers of gay men (13/97=13%) compared to controls (4/103=4%) and increased in mothers with two or more gay sons (10/44=23%). Our findings support a role for the X chromosome in regulating sexual orientation in a subgroup of gay men.

Case-Control Studies↗

PubMed2005

Targeted disruption of glycerol kinase gene in mice: expression analysis in liver shows alterations in network partners related to glycerol kinase activity.

Glycerol kinase deficiency (GKD) is an X-linked inborn error of metabolism with metabolic and neurological crises. Liver shows the highest level of glycerol kinase (GK) activity in humans and mice. Absence of genotype-phenotype correlations in patients with GKD indicates the involvement of modifier genes, including other network partners. To understand the molecular pathogenesis of GKD, we performed microarray analysis on liver mRNA from neonatal glycerol kinase (Gyk) knockout (KO) and wild-type (WT) mice. Unsupervised learning revealed that the overall gene expression profile of the KO mice was different from that of WT. Real-time PCR confirmed the differences for selected genes. Functional gene enrichment analysis was used to find 56 increased and 37 decreased gene functional categories. PathwayAssist analysis identified changes in gene expression levels of genes involved in organic acid metabolism indicating that GK was part of the same metabolic network which correlates well with the patients with GKD having metabolic acidemia during their episodic crises. Network component analysis (NCA) showed that transcription factors sterol regulatory element-binding protein (SREBP)-1c, carbohydrate response element-binding protein (ChREBP), hepatocyte nuclear factor-4 alpha (HNF-4alpha) and peroxisome proliferative-activated receptor-alpha (PPARalpha) had increased activity in the Gyk KO mice compared with WT mice, whereas SREBP-2 was less active in the Gyk KO mice. These studies show that Gyk deletion causes alterations in expression of genes in several regulatory networks and is the first time NCA has been used to expand on microarray data from a mouse KO model of a human disease.

Animals↗

PubMed2005

Distinct gene expression profiles in adult mouse heart following targeted MAP kinase activation.

Three major MAP kinase signaling cascades, ERK, p38, and JNK, play significant roles in the development of cardiac hypertrophy and heart failure in response to external stress and neural/hormonal stimuli. To study the specific function of each MAP kinase branch in adult heart, we have generated three transgenic mouse models with cardiac-specific and temporally regulated expression of activated mutants of Ras, MAP kinase kinase (MKK)3, and MKK7, which are selective upstream activators for ERK, p38, and JNK, respectively. Gene expression profiles in transgenic adult hearts were determined using cDNA microarrays at both early (4-7 days) and late (2-4 wk) time points following transgene induction. From this study, we revealed common changes in gene expression among the three models, particularly involving extracellular matrix remodeling. However, distinct expression patterns characteristic for each pathway were also identified in cell signaling, growth, and physiology. In addition, genes with dynamic expression differences between early vs. late stages illustrated primary vs. secondary changes on MAP kinase activation in adult hearts. These results provide an overview to both short-term and long-term effects of MAP kinase activation in heart and support some common as well as unique roles for each MAP kinase cascade in the development of heart failure.

Animals↗

PubMed2005

Molecular determinants of the response of glioblastomas to EGFR kinase inhibitors.

BACKGROUND: The epidermal growth factor receptor (EGFR) is frequently amplified, overexpressed, or mutated in glioblastomas, but only 10 to 20 percent of patients have a response to EGFR kinase inhibitors. The mechanism of responsiveness of glioblastomas to these inhibitors is unknown. METHODS: We sequenced kinase domains in the EGFR and human EGFR type 2 (Her2/neu) genes and analyzed the expression of EGFR, EGFR deletion mutant variant III (EGFRvIII), and the tumor-suppressor protein PTEN in recurrent malignant gliomas from patients who had received EGFR kinase inhibitors. We determined the molecular correlates of clinical response, validated them in an independent data set, and identified effects of the molecular abnormalities in vitro. RESULTS: Of 49 patients with recurrent malignant glioma who were treated with EGFR kinase inhibitors, 9 had tumor shrinkage of at least 25 percent. Pretreatment tissue was available for molecular analysis from 26 patients, 7 of whom had had a response and 19 of whom had rapid progression during therapy. No mutations in EGFR or Her2/neu kinase domains were detected in the tumors. Coexpression of EGFRvIII and PTEN was significantly associated with a clinical response (P<0.001; odds ratio, 51; 95 percent confidence interval, 4 to 669). These findings were validated in 33 patients who received similar treatment for glioblastoma at a different institution (P=0.001; odds ratio, 40; 95 percent confidence interval, 3 to 468). In vitro, coexpression of EGFRvIII and PTEN sensitized glioblastoma cells to erlotinib. CONCLUSIONS: Coexpression of EGFRvIII and PTEN by glioblastoma cells is associated with responsiveness to EGFR kinase inhibitors.

Adult↗

PubMed2005

Inhibition of Atm and/or Atr disrupts gene silencing on the inactive X chromosome.

ATM and ATR are well documented for their roles in maintaining the integrity of genomic DNA by responding to DNA damage and preparing the cell for repair. Since ATM and ATR have been reported to exist in complexes with histone deacetylases, we asked whether Atm and Atr might also uphold gene silencing by heterochromatin. We show that the Atm/Atr inhibitor 2-aminopurine causes the inactive X chromosome to accumulate abnormal chromatin and undergo unwanted gene reactivation. We provide evidence that this gene expression from the inactive X chromosome is not a byproduct of the accumulation of DNA breaks. Individually inhibiting Atm and Atr by either small interfering RNA or the expression of dominant-negative ATM and ATR constructs also compromised X-inactivation. Atm and Atr, therefore, not only function in responding to DNA damage but perhaps also are involved in gene silencing via the maintenance of heterochromatin.

2-Aminopurine↗

PubMed2005

A general framework for weighted gene co-expression network analysis.

Gene co-expression networks are increasingly used to explore the system-level functionality of genes. The network construction is conceptually straightforward: nodes represent genes and nodes are connected if the corresponding genes are significantly co-expressed across appropriately chosen tissue samples. In reality, it is tricky to define the connections between the nodes in such networks. An important question is whether it is biologically meaningful to encode gene co-expression using binary information (connected=1, unconnected=0). We describe a general framework for ;soft' thresholding that assigns a connection weight to each gene pair. This leads us to define the notion of a weighted gene co-expression network. For soft thresholding we propose several adjacency functions that convert the co-expression measure to a connection weight. For determining the parameters of the adjacency function, we propose a biologically motivated criterion (referred to as the scale-free topology criterion). We generalize the following important network concepts to the case of weighted networks. First, we introduce several node connectivity measures and provide empirical evidence that they can be important for predicting the biological significance of a gene. Second, we provide theoretical and empirical evidence that the ;weighted' topological overlap measure (used to define gene modules) leads to more cohesive modules than its ;unweighted' counterpart. Third, we generalize the clustering coefficient to weighted networks. Unlike the unweighted clustering coefficient, the weighted clustering coefficient is not inversely related to the connectivity. We provide a model that shows how an inverse relationship between clustering coefficient and connectivity arises from hard thresholding. We apply our methods to simulated data, a cancer microarray data set, and a yeast microarray data set.

Journal Article↗

PubMed2005

Global histone modification patterns predict risk of prostate cancer recurrence.

Aberrations in post-translational modifications of histones have been shown to occur in cancer cells but only at individual promoters; they have not been related to clinical outcome. Other than being targeted to promoters, modifications of histones, such as acetylation and methylation of lysine and arginine residues, also occur over large regions of chromatin including coding regions and non-promoter sequences, which are referred to as global histone modifications. Here we show that changes in global levels of individual histone modifications are also associated with cancer and that these changes are predictive of clinical outcome. Through immunohistochemical staining of primary prostatectomy tissue samples, we determined the percentage of cells that stained for the histone acetylation and dimethylation of five residues in histones H3 and H4. Grouping of samples with similar patterns of modifications identified two disease subtypes with distinct risks of tumour recurrence in patients with low-grade prostate cancer. These histone modification patterns were predictors of outcome independently of tumour stage, preoperative prostate-specific antigen levels, and capsule invasion. Thus, widespread changes in specific histone modifications indicate previously undescribed molecular heterogeneity in prostate cancer and might underlie the broad range of clinical behaviour in cancer patients.

Acetylation↗

PubMed2005

Clinicopathologic and molecular correlations of necrosis in the primary tumor of patients with renal cell carcinoma.

BACKGROUND: The presence of histologic necrosis in the primary tumor of patients with renal cell carcinoma (RCC) has been suggested to be an important predictor of survival. The authors investigated the relation of tumor necrosis to other clinicopathologic factors known to be important prognostic indicators for patients with RCC. METHODS: The records of 311 patients undergoing treatment for RCC were evaluated for basic clinicopathologic information including TNM classification, nuclear grade, Eastern Cooperative Oncology Group (ECOG) performance status (PS), disease recurrence, and survival. The presence and extent of histologic necrosis of the primary tumors was recorded and correlated with clinicopathologic factors, carbonic anhydrase IX and Ki-67 expression, disease recurrence, and survival. RESULTS: The presence of necrosis in the primary tumor of patients with RCC compared with patients with RCC without necrosis was associated with higher T classification (P < 0.0001), the presence of lymph node disease (P = 0.009), the presence of metastases (P < 0.0001), higher grade (P < 0.0001), greater mean tumor size (P < 0.0001), an ECOG PS score > or = 1 (P = 0.007), higher University of California-Los Angeles Integrated Staging System (UISS) category (P < 0.0001), and higher Ki-67 expression (P < 0.0001). The extent of necrosis in the primary tumor was associated with the presence of lymph node disease (P = 0.009) and the presence of metastases (P < 0.0001), and correlated with higher T classification (sigma = 0.31, P < 0.0001), poorer ECOG PS (sigma = 0.18, P = 0.002), higher grade (sigma = 0.33, P < 0.0001), greater tumor size (sigma = 0.40, P < 0.0001), higher UISS category (sigma = 0.37, P < 0.0001), and higher Ki-67 staining (sigma = 0.32, P < 0.0001). Patients with the presence of necrosis in the primary tumor demonstrated a lower 5-year disease-specific survival compared with patients without necrosis in the primary tumor (36% vs. 75%; P < 0.0001). Multivariate analysis demonstrated that T classification (P < 0.0001), distant metastases (P < 0.0001), and ECOG PS (P < 0.0001) were independent predictors of DSS, whereas the presence of necrosis was not (P = 0.1100). Substratification into localized and metastatic disease demonstrated that the presence of necrosis was an independent predictor of survival in patients with localized (P = 0.025), but not metastatic (P = 0.44), disease. The extent of necrosis was not an independent predictor of survival (P > 0.05). Patients with the presence of necrosis in the primary tumor had a lower 5-year disease recurrence-free rate compared with patients without the presence of necrosis (62% vs. 92%, P < 0.0001). CONCLUSIONS: The presence of necrosis in the primary tumor was associated with adverse prognostic factors such as high T classification, presence of lymph node disease and metastases, high grade, large tumor size, and poor ECOG PS. The extent of necrosis was found to be associated with the presence of lymph node disease and metastases and correlated with higher T classification, higher grade, greater tumor size, poorer ECOG PS, and higher UISS category. The presence of this histologic variant was an independent predictor of poor survival in patients with localized, but not metastatic, disease. In addition, Ki-67 expression served as a valuable surrogate marker for the presence of histologic tumor necrosis.

Adenocarcinoma, Clear Cell↗