Armin O. Schmitt

Finite sample effects in sequence analysis

Abstract This paper is devoted to the statistical analysis of symbol sequences, such as Markov strings, DNA sequences, or texts from natural languages. It is shown that entropy calculations are seriously affected by systematic errors due to the finite size of the samples. These difficulties can be dealt with by assuming simple probability distributions underlying the generating process (e.g. equidistribution, power-law distribution, exponential distribution). Analytical expressions for the dominant correction terms are derived and tested.

A New Method to Calculate Higher-Order Entropies from Finite Samples

The finite size of natural symbol sequences, e.g. DNA strands or texts in natural languages, is responsible for the worsening statistics with respect to increasing length n of the substrings, thus restricting the reliability of the results of higher-order entropy calculations to small n. A new method for the calculation of higher-order entropies Hn based upon a theorem of coding theory is presented, allowing for reliable estimations far beyond. We tested the range of validity of this method by means of symbol sequences with known entropies: two stochastic processes (the underlying probability distribution being the equidistribution and the distribution of the nucleotides of the yeast chromosome III DNA sequence) and a Markov process of fifth order with its transition probabilities also taken from this yeast DNA sequence.

THE MODULAR STRUCTURE OF INFORMATIONAL SEQUENCES

It is shown that DNA sequences can be decomposed into smaller units much the same as texts can be decomposed into syllables, words, or groups of words. Those smaller units (modules) are extracted from DNA sequences according to statistical criteria. Tests with sequences of known modular structure (two novels and a FORTRAN source code) were performed. The rate to which DNA sequences can be decomposed into modules (modularity) turns out to be a very sensitive measure to distinguish DNA sequences from random sequences.

Early and late onset of voluntary exercise have differential effects on the metabolic syndrome in an obese mouse model.

In a mouse model for juvenile obesity, we investigated how the age of onset of voluntary exercise affects factors of the metabolic syndrome. One exercise group had access to running wheels from 3 weeks (representing childhood) and another one from 9 weeks on (early adulthood). Both groups were compared to mice without exercise. The investigations were performed under 2 diets (standard maintenance and high-fat diet). Average daily running activity was independent of diet and exercise. On both diets, mice with exercise from 3 weeks on gained 10 g body weight and 5 g fat mass less than mice without exercise. The highest body weight difference between mice on HFD without exercise and mice on standard maintenance diet with exercise was 24 g. Despite the higher energy expenditure during exercise, young mice did not increase their energy intake adjusted for lean mass, while mice with exercise from 9 weeks had an increased energy intake of 6 kJ per day and therefore could not reduce fat mass on both diets. However, mice with exercise from 9 weeks had better glucose tolerance at 20 weeks than mice with exercise from childhood on. Independently of the age of exercise onset, triglycerides were reduced from 2.4 to 1.7 mmol/l on both diets and insulin levels from 1.5 to 0.3 and 4.5 to 1.8 µg/ml on standard maintenance and high-fat diet, respectively, which represents a considerable improvement. Physical activity seems to have long-lasting effects on body composition and health, but they are different depending on when exercise has begun.

High-density cDNA Grids for hybridization fingerprinting experiments

Publisher Summary This chapter discusses the use of high-density cDNA grids for hybridization fingerprinting experiments. Hybridization fingerprinting experiments with short oligonucleotides as probes provide a method to screen efficiently at relatively low costs without losing much information. It could be shown that the analysis of whole genomes with this technique is viable using automated protocols at various steps. The chapter presents a protocol for the production of high-density grids for the purpose. It explores a number of built-in controls to judge their quality. And a number of possible applications are also listed here. The protocol outlines construction of cDNA libraries, PCR amplification, and production of high-density cDNA grids. The discussion closes with a conclusion that despite some shortcomings of the technique, hybridization fingerprinting is a powerful method to very efficiently extract valuable information about large quantities of cDNA material. Attaining a survey of the mRNA expression pattern of a cell in as little as 6 weeks (including production of library, hybridization, and data analysis) is an achievement which is unrivaled by other experimental approaches.

Where in the genome are significant single nucleotide polymorphisms from genome-wide association studies located?

Recent technological progress has permitted the efficient performance of genome-wide association studies (GWAS) to map genetic variants associated with common diseases. Here, we analyzed 2,893 single nucleotide polymorphisms (SNPs) that have been identified in 593 published GWAS as associated with a disease phenotype with respect to their genomic location. In absolute numbers, most significant SNPs are located in intergenic regions and introns. When compared to their representation on the chips, there is essentially overrepresentation of nonsynonymous coding SNPs (nsSNPs), synonymous coding SNPs, and SNPs in untranscribed regions upstream of genes among the disease associated SNPs. A Gene Ontology term analysis showed that genes putatively causing a phenotype often code for membrane associated proteins or signal transduction genes.

Genetic determinants for intramuscular fat content and water-holding capacity in mice selected for high muscle mass

Intramuscular fat content and water-holding capacity are important traits in livestock as they influence meat quality, nutritive value of the muscle, and animal health. As a model for livestock, two inbred lines of the Berlin Muscle Mouse population, which had been long-term selected for high muscle mass, were used to identify genomic regions affecting intramuscular fat content and water-holding capacity. The intramuscular fat content of the Musculus longissimus was on average 1.4 times higher in BMMI806 than in BMMI816 mice. This was accompanied by a 1.5 times lower water-holding capacity of the Musculus quadriceps in BMMI816 mice. Linkage analyses with 332 G3 animals of reciprocal crosses between these two lines revealed quantitative trait loci for intramuscular fat content on chromosome 7 and for water-holding capacity on chromosome 2. In part, the identified loci coincide with syntenic regions in pigs in which genetic effects for the same traits were found. Therefore, these muscle-weight-selected mouse lines and the produced intercross populations are valuable genetic resources to identify genes that could also contribute to meat quality in other species.

RandoMate: a program for the generation of random mating schemes for small laboratory animals.

Advanced intercross lines (AIL) have proven to be a powerful tool in genetic research to map complex genetic traits. The advantage of AIL is the high enrichment of visible recombination events to fine map the position of the target gene. Therefore, AIL are generated under the avoidance of inbreeding. We developed an online software tool, RandoMate, that generates random mating schemes such that only animals from different families are paired. When animals have to be selected randomly for mating, RandoMate optimizes the mating scheme such that all families contribute equally to the next generation. RandoMate uses a divide-and-conquer algorithm to define a mating scheme without brother-sister matings for all animals of a generation. If not all animals can be considered for the next generation, the mating scheme maximizes the randomness of the occurrences of animals from their families to make the family contributions as equal as possible. RandoMate is freely available at http://www2.hu-berlin.de/RandoMate.

Genetic and diet effects on Ppar-α and Ppar-γ signaling pathways in the Berlin Fat Mouse Inbred line with genetic predisposition for obesity

BackgroundThe Berlin Fat Mouse Inbred (BFMI) line is a new mouse model for obesity, which was long-term selected for high fatness. Peroxisome proliferator-activated receptors (PPARs) are involved in the control of energy homeostasis, nutrient metabolism and cell proliferation. Here, we studied the expression patterns of the different Ppar genes and the genes in the PPAR pathway in the BFMI line in comparison to physiological changes.ResultsAt the age of 10 weeks, the BFMI mice exhibited marked obesity with enlarged adipocytes and high serum triglycerides concentrations in comparison to the often used mouse line C57BL/6 (B6). Between these two lines, gene expression analyses revealed differentially expressed genes belonging to the PPAR pathway, in particular genes of the lipogenesis and the fatty acid transport.ConclusionSurprisingly, the Ppar-α gene expression was up-regulated in liver and Ppar-γ gene expression was down-regulated in the white adipose tissue, indicating the activation of a mechanism that counteracts the rise of obesity.

Tracking chromosomal positions of oligomers - a case study with Illumina's BovineSNP50 beadchip

BackgroundHigh density genotyping arrays have become established as a valuable research tool in human genetics. Currently, more than 300 genome wide association studies were published for human reporting about 1,000 SNPs that are associated with a phenotype. Also in animal sciences high density genotyping arrays are harnessed to analyse genetic variation. To exploit the full potential of this technology single nucleotide polymorphisms (SNPs) on the chips should be well characterized and their chromosomal position should be precisely known. This, however, is a challenge if the genome sequence is still subject to changes.ResultsWe have developed a mapping strategy and a suite of software scripts to update the chromosomal positions of oligomer sequences used for SNP genotyping on high density arrays. We describe the mapping procedure in detail so that scientists with moderate bioinformatics skills can reproduce it. We furthermore present a case study in which we re-mapped 54,001 oligomer sequences from Iluminas BovineSNP50 beadchip to the bovine genome sequence. We found in 992 cases substantial discrepancies between the manufacturers annotations and our results. The software scripts in the Perl and R programming languages are provided as supplements.ConclusionsThe positions of oligomer sequences in the genome are volatile even within one build of the genome. To facilitate the analysis of data from a GWAS or from an expression study, especially with species whose genome assembly is still unstable, it is recommended to update the oligomer positions before data analysis.