8 December 2003 in Proceedings of the National Academy of Sciences
Many stochastic simulation approaches for generating observations from a posterior distribution depend on knowing a likelihood function. However, for many complex probability models, such likelihoods are either impossible or computationally prohibitive to obtain. Here we present a Markov chain Monte Carlo method for generating observations from a posterior distribution without the use of likelihoods. It can also be used in frequentist applications, in particular for maximum-likelihood estimation. The approach is illustrated by an example of ancestral inference in population genetics. A number of open problems are highlighted in the discussion.
26 October 2006 in BMC Genomics
Background: Eukaryotic replication origins exhibit different initiation efficiencies and activation times within S-phase. Although local chromatin structure and function influences origin activity, the exact mechanisms remain poorly understood. A key to understanding the exact features of chromatin that impinge on replication origin function is to define the precise locations of the DNA sequences that control origin function. InS. cerevisiae, Autonomously Replicating Sequences (ARSs) contain a consensus sequence (ACS) that binds the Origin Recognition Complex (ORC) and is essential for origin function. However, an ACS is not sufficient for origin function and the majority of ACS matches do not function as ORC binding sites, complicating the specific identification of these sites.Results: To identify essential origin sequences genome-wide, we utilized a tiled oligonucleotide array (NimbleGen) to map the ORC and Mcm2p binding sites at high resolution. These binding sites define a set of potential Autonomously Replicating Sequences (ARSs), which we term nimARSs. The nimARS set comprises 529 ORC and/or Mcm2p binding sites, which includes 95% of known ARSs, and experimental verification demonstrates that 94% are functional. The resolution of the analysis facilitated identification of potential ACSs (nimACSs) within 370 nimARSs. Cross-validation shows that the nimACS predictions include 58% of known ACSs, and experimental verification indicates that 82% are essential for ARS activity.Conclusion: These findings provide the most comprehensive, accurate, and detailed mapping of ORC binding sites to date, adding to the emerging picture of the chromatin organization of the budding yeast genome.
9 December 2007 in Genome Biology
Background: Several interventions increase lifespan in model organisms, including reduced insulin/insulin-like growth factor-like signaling (IIS), FOXO transcription factor activation, dietary restriction, and superoxide dismutase (SOD) over-expression. One question is whether these manipulations function through different mechanisms, or whether they intersect on common processes affecting aging. Results: A doxycycline-regulated system was used to over-express manganese-SOD (MnSOD) in adult Drosophila, yielding increases in mean and maximal lifespan of 20%. Increased lifespan resulted from lowered initial mortality rate and required MnSOD over-expression in the adult. Transcriptional profiling indicated that the expression of specific genes was altered by MnSOD in a manner opposite to their pattern during normal aging, revealing a set of candidate biomarkers of aging enriched for carbohydrate metabolism and electron transport genes and suggesting a true delay in physiological aging, rather than a novel phenotype. Strikingly, cross-dataset comparisons indicated that the pattern of gene expression caused by MnSOD was similar to that observed in long-lived Caenorhabditis elegans insulin-like signaling mutants and to the xenobiotic stress response, thus exposing potential conserved longevity promoting genes and implicating detoxification in Drosophila longevity. Conclusion: The data suggest that MnSOD up-regulation and a retrograde signal of reactive oxygen species from the mitochondria normally function as an intermediate step in the extension of lifespan caused by reduced insulin-like signaling in various species. The results implicate a species-conserved net of coordinated genes that affect the rate of senescence by modulating energetic efficiency, purine biosynthesis, apoptotic pathways, endocrine signals, and the detoxification and excretion of metabolites.
23 June 2005 in BMC Medicine
Background: The small intestine is constructed of many crypts and villi, and mouse studies suggest that each crypt contains multiple stem cells. Very little is known about human small intestines because mouse fate mapping strategies are impractical in humans. However, it is theoretically possible that stem cell histories are inherently written within their genomes. Genomes appear to record histories (as exemplified by use of molecular clocks), and therefore it may be possible to reconstruct somatic cell dynamics from somatic cell errors. Recent human colon studies suggest that random somatic epigenetic errors record stem cell histories (ancestry and total numbers of divisions). Potentially age-related methylation also occurs in human small intestines, which would allow characterization of their stem cells and comparisons with the colon. Methods: Methylation patterns in individual crypts from 13 small intestines (17 to 78 years old) were measured by bisulfite sequencing. The methylation patterns were analyzed by a quantitative model to distinguish between immortal or niche stem cell lineages. Results: Age-related methylation was observed in the human small intestines. Crypt methylation patterns were more consistent with stem cell niches than immortal stem cell lineages. Human large and small intestine crypt niches appeared to have similar stem cell dynamics, but relatively less methylation accumulated with age in the small intestines. There were no apparent stem cell differences between the duodenum and ileum, and stem cell survival did not appear to decline with aging. Conclusion: Crypt niches containing multiple stem cells appear to maintain human small intestines. Crypt niches appear similar in the colon and small intestine, and the small intestinal stem cell mitotic rate is the same as or perhaps slower than that of the colon. Although further studies are needed, age-related methylation appears to record somatic cell histories, and a somatic epigenetic molecular clock strategy may potentially be applied to other human tissues to reconstruct otherwise occult stem cell histories.