Interactions between natural selection, recombination and gene density in the genes of Drosophila

In Drosophila, as in many organisms, natural selection leads to high levels of codon bias in genes that are highly expressed. Thus codon bias is an indicator of the intensity of one kind of selection that is experienced by genes and can be used to assess the impact of other genomic factors on natural selection. Among 13,000 genes in the Drosophila genome, codon bias has a slight positive, and strongly significant, association with recombination--as expected if recombination allows natural selection to act more efficiently when multiple linked sites segregate functional variation. The same reasoning leads to the expectation that the efficiency of selection, and thus average codon bias, should decline with gene density. However, this prediction is not confirmed. Levels of codon bias and gene expression are highest for those genes in an intermediate range of gene density, a pattern that may be the result of a tradeoff between the advantages for gene expression of close gene spacing and disadvantages arising from regulatory conflicts among tightly packed genes. These factors appear to overlay the more subtle effect of linkage among selected sites that gives rise to the association between recombination rate and codon bias.     

The localization and expression of the class II starch synthases of wheat.

The starch granules of hexaploid wheat (Triticum aestivum) contain a group of three proteins known as SGP-1 (starch granule protein-1) proteins, which have apparent molecular masses of 100, 108, and 115 kD. The nature and role of these proteins has not been defined previously. We demonstrate that these polypeptides are starch synthases that are present in both the starch granule and the soluble fraction at the early stages of wheat endosperm development, but that are exclusively granule bound at mid and late endosperm development. A partial cDNA clone encoding a fragment of the 100-kD protein was obtained by screening a wheat endosperm cDNA expression library using monoclonal antibodies. Three classes of cDNA were subsequently isolated from a wheat endosperm cDNA library by nucleic acid hybridization and were shown to encode the 100-, 108-, and 115-kD proteins. The cDNA sequences are highly homologous to class II starch synthases and have the highest homology with the maize SSIIa (starch synthase IIa) gene. mRNA for the SGP-1 proteins was detected in the leaf, pre-anthesis florets, and endosperm of wheat and is highly expressed in the leaf and in the grain during the early to mid stages of development. We discuss the roles of the SGP-1 proteins in starch biosynthesis in wheat.     

Estimating Divergence Parameters With Small Samples From a Large Number of Loci

Most methods for studying divergence with gene flow rely upon data from many individuals at few loci. Such data can be useful for inferring recent population history but they are unlikely to contain sufficient information about older events. However, the growing availability of genome sequences suggests a different kind of sampling scheme, one that may be more suited to studying relatively ancient divergence. Data sets extracted from whole-genome alignments may represent very few individuals but contain a very large number of loci. To take advantage of such data we developed a new maximum-likelihood method for genomic data under the isolation-with-migration model. Unlike many coalescent-based likelihood methods, our method does not rely on Monte Carlo sampling of genealogies, but rather provides a precise calculation of the likelihood by numerical integration over all genealogies. We demonstrate that the method works well on simulated data sets. We also consider two models for accommodating mutation rate variation among loci and find that the model that treats mutation rates as random variables leads to better estimates. We applied the method to the divergence of Drosophila melanogaster and D. simulans and detected a low, but statistically significant, signal of gene flow from D. simulans to D. melanogaster.IN the study of speciation researchers often inquire of the extent that populations have exchanged genes as they diverged and on the time since populations began to diverge. Answers to questions about historical divergence and gene flow potentially lie in patterns of genetic variation that are found in present day populations. To bridge the gap between population history and current genetic data, population geneticists can make use of a gene genealogy, G, a bifurcating tree that represents the history of ancestry of sampled gene copies. The probability of a particular value of G can be calculated for a particular parameter set using coalescent models. Then given a particular genealogy, genetic variation can be examined using a mutation model that is appropriate for the kind of data being used. Finally by considering multiple values of G, the connection can be made between the population evolution history and the data. A mathematical representation that treats G as a key interstitial variable was given by Felsenstein (1988),(1)where X represents the sequence data, G represents gene genealogy, Ψ represents the set of all possible genealogies, and Θ represents the vector of population parameters included in the model.Unless sample sizes are very small, (1) cannot be solved analytically, and so considerable effort has gone into finding approximate solutions (Kuhner et al. 1995; Griffiths and Marjoram 1996; Wilson and Balding 1998). One general approach is to sample genealogies using a Markov chain Monte Carlo (MCMC) simulation. This is the approach developed by Kuhner and colleagues (Kuhner et al. 1995) and that has since been extended to models with migration (Beerli and Felsenstein 1999, 2001; Nielsen and Wakeley 2001). A general problem for these methods is that they usually require long running times to generate sufficiently large and independent samples, especially when the MCMC simulation is mixing slowly.With fast-improving DNA sequencing techniques, more and more genome sequences are becoming available, and alignments of these whole-genome sequences are a very useful source of information for the study of divergence. However, traditional MCMC methods are likely to be slow on genome-scale data because running times are proportional to the number of loci. To overcome this difficulty Yang developed a likelihood method (Yang 2002) for data sets containing one sample from each of the three populations at every locus. This method uses numerical integration to calculate the likelihood function in Equation 1. By using a very large number of loci, the method can make up for using a very small number of individuals (i.e., genomes).Yang''s method is based on a divergence model that assumes no gene flow between separated populations. However, there are many situations where gene flow may have been occurring and where it is preferable to include it within the divergence model. One model that has been used a lot in this context is the isolation-with-migration (IM) model, which incorporates both population separation and migration (Nielsen and Wakeley 2001). Under an IM model the genealogies include not only some fixed number of coalescent events and speciation events, but also any possible number of migration events. The potential for very large numbers of migration events complicates the sample space of G and makes the numerical integration seemingly impossible. Innan and Watanabe (2006) circumvent this problem by using a recursion method to estimate the coalescent rates on a series of time points. In their recursion, the accuracy in calculating coalescent rate at one time point depends on the accuracy of calculations at previous time points, and this may impair the precision of the overall likelihood calculation. Therefore we developed a method that relies on numerical integration to calculate the likelihood under an IM model. We tested the accuracy of this method on simulated data sets of various sample sizes and applied it to a genome alignment of Drosophila melanogaster and D. simulans (with D. yakuba as an outgroup).     

Non-equilibrium allele frequency spectra via spectral methods

A major challenge in the analysis of population genomics data consists of isolating signatures of natural selection from background noise caused by random drift and gene flow. Analyses of massive amounts of data from many related populations require high-performance algorithms to determine the likelihood of different demographic scenarios that could have shaped the observed neutral single nucleotide polymorphism (SNP) allele frequency spectrum. In many areas of applied mathematics, Fourier Transforms and Spectral Methods are firmly established tools to analyze spectra of signals and model their dynamics as solutions of certain Partial Differential Equations (PDEs). When spectral methods are applicable, they have excellent error properties and are the fastest possible in high dimension; see Press et al. (2007). In this paper we present an explicit numerical solution, using spectral methods, to the forward Kolmogorov equations for a Wright–Fisher process with migration of K populations, influx of mutations, and multiple population splitting events.     

DNA sequences from Arabidopsis, which encode protein kinases and function as upstream regulators of Snf1 in yeast

Sucrose nonfermenting-1 (Snf1)-related protein kinase-1 (SnRK1) of plants is a global regulator of carbon metabolism through the modulation of enzyme activity and gene expression. It is structurally and functionally related to the yeast protein kinase, Snf1, and to mammalian AMP-activated protein kinase. Two DNA sequences from Arabidopsis thaliana, previously known only by their data base accession numbers of NM_ 125448.3 (protein ID NP_200863) and NM_114393.3 (protein ID NP_566876) each functionally complemented a Saccharomyces cerevisiae elm1 sak1 tos3 triple mutant. This indicates that the Arabidopsis proteins are able to substitute for one of the missing yeast upstream kinases, which are required for activity of Snf1. Both plant proteins were shown to phosphorylate a peptide with the amino acid sequence of the phosphorylation site in the T-loop of SnRK1 and by inference SnRK1 in Arabidopsis. The proteins encoded by NM_125448.3 and NM_114393.3 have been named AtSnAK1 and AtSnAK2 (Arabidopsis thaliana SnRK1-activating kinase), respectively. We believe this is the first time that upstream activators of SnRK1 have been described in any plant species.