Inferring the deep past from molecular data

There is an expectation that analyses of molecular sequences might be able to distinguish between alternative hypotheses for ancient relationships, but the phylogenetic methods used and types of data analyzed are of critical importance in any attempt to recover historical signal. Here we discuss some common issues that can influence the topology of trees obtained when using overly-simple models to analyze molecular data that often display complicated patterns of sequence heterogeneity. To illustrate our discussion, we have used three examples of inferred relationships which have changed radically as models and methods of analysis have improved. In two of these examples, the sister-group relationship between thermophilic Thermus and mesophilic Deinococcus, and the position of long-branch Microsporidia among eukaryotes, we show that recovering what is now generally considered to be the correct tree is critically dependent on the fit between model and data. In the third example, the position of eukaryotes in the tree of life, the hypothesis that is currently supported by the best available methods is fundamentally different from the classical view of relationships between major cellular domains. Since heterogeneity appears to be pervasive and varied among all molecular sequence data, and even the best available models can still struggle to deal with some problems, the issues we discuss are generally relevant to phylogenetic analyses. It remains essential to maintain a critical attitude to all trees as hypotheses of relationship that may change with more data and better methods.

A Novel Test for Absolute Fit of Evolutionary Models Provides a Means to Correctly Identify the Substitution Model and the Model Tree

A novel test is described that visualizes the absolute model-data fit of the substitution and tree components of an evolutionary model. The test utilizes statistics based on counts of character state matches and mismatches in alignments of observed and simulated sequences. This comparison is used to assess model-data fit. In simulations conducted to evaluate the performance of the test, the test estimator was able to identify both the correct tree topology and substitution model under conditions where the Goldman–Cox test—which tests the fit of a substitution model to sequence data and is also based on comparing simulated replicates with observed data—showed high error rates. The novel test was found to identify the correct tree topology within a wide range of DNA substitution model misspecifications, indicating the high discriminatory power of the test. Use of this test provides a practical approach for assessing absolute model-data fit when testing phylogenetic hypotheses.

Four-derivative couplings and BPS dyons in heterotic CHL orbifolds

Three-dimensional string models with half-maximal supersymmetry are believed to be invariant under a large U-duality group which unifies the S and T dualities in four dimensions. We propose an exact, U-duality invariant formula for four-derivative scalar couplings of the form F(\Phi) (\nabla\Phi)^4F(Φ)(∇Φ)4 in a class of string vacua known as CHL \IZ_N heterotic orbifolds with NN prime, generalizing our previous work which dealt with the case of heterotic string on T^6T6. We derive the Ward identities that F(\Phi)F(Φ) must satisfy, and check that our formula obeys them. We analyze the weak coupling expansion of F(\Phi)F(Φ), and show that it reproduces the correct tree-level and one-loop contributions, plus an infinite series of non-perturbative contributions. Similarly, the large radius expansion reproduces the exact F^4F4 coupling in four dimensions, including both supersymmetric invariants, plus infinite series of instanton corrections from half-BPS dyons winding around the large circle, and from Taub-NUT instantons. The summation measure for dyonic instantons agrees with the helicity supertrace for half-BPS dyons in 4 dimensions in all charge sectors. In the process we clarify several subtleties about CHL models in D=4D=4 and D=3D=3, in particular we obtain the exact helicity supertraces for 1/2-BPS dyonic states in all duality orbits.

StrainSeeker: fast identification of bacterial strains from raw sequencing reads using user-provided guide trees

Background Fast, accurate and high-throughput identification of bacterial isolates is in great demand. The present work was conducted to investigate the possibility of identifying isolates from unassembled next-generation sequencing reads using custom-made guide trees. Results A tool named StrainSeeker was developed that constructs a list of specific k-mers for each node of any given Newick-format tree and enables the identification of bacterial isolates in 1–2 min. It uses a novel algorithm, which analyses the observed and expected fractions of node-specific k-mers to test the presence of each node in the sample. This allows StrainSeeker to determine where the isolate branches off the guide tree and assign it to a clade whereas other tools assign each read to a reference genome. Using a dataset of 100 Escherichia coli isolates, we demonstrate that StrainSeeker can predict the clades of E. coli with 92% accuracy and correct tree branch assignment with 98% accuracy. Twenty-five thousand Illumina HiSeq reads are sufficient for identification of the strain. Conclusion StrainSeeker is a software program that identifies bacterial isolates by assigning them to nodes or leaves of a custom-made guide tree. StrainSeeker’s web interface and pre-computed guide trees are available at http://bioinfo.ut.ee/strainseeker. Source code is stored at GitHub: https://github.com/bioinfo-ut/StrainSeeker.

TreeToReads - a pipeline for simulating raw reads from phylogenies

Using genome-wide SNP-based methods for tracking pathogens has become standard practice in academia and public health agencies. There are multiple computational approaches available that perform a similar task: call variants by mapping short read data against a reference genome, quality filter these variants, then concatenate the variants into a sequence matrix for downstream phylogenetic analysis. However, there are no existing methods to validate the accuracy of these approaches despite the fact that we know there are parameters that can affect whether a SNP is called, or the correct tree is recovered. We present a simulation approach (TreeToReads) to generate raw read data from mutated genomes simulated under a known phylogeny. The user can vary parameters of interest at each step in the simulation (e.g., topology, model of sequence evolution, and read coverage) to assess the robustness of a given result, which is critical within both research and applied settings. Source code, examples, and a tutorial are available at \url{https://github.com/snacktavish/TreeToReads}.

Inferring tree causal models of cancer progression with probability raising

Existing techniques to reconstruct tree models of progression for accumulative processes, such as cancer, seek to estimate causation by combining correlation and a frequentist notion of temporal priority. In this paper, we define a novel theoretical framework called CAPRESE (CAncer PRogression Extraction with Single Edges) to reconstruct such models based on the notion of probabilistic causation defined by Suppes. We consider a general reconstruction setting complicated by the presence of noise in the data due to biological variation, as well as experimental or measurement errors. To improve tolerance to noise we define and use a shrinkage-like estimator. We prove the correctness of our algorithm by showing asymptotic convergence to the correct tree under mild constraints on the level of noise. Moreover, on synthetic data, we show that our approach outperforms the state-of-the-art, that it is efficient even with a relatively small number of samples and that its performance quickly converges to its asymptote as the number of samples increases. For real cancer datasets obtained with different technologies, we highlight biologically significant differences in the progressions inferred with respect to other competing techniques and we also show how to validate conjectured biological relations with progression models.

(327) Correcting Ni Deficiency in Pecan and Other Crops

The discovery of nickel (Ni) deficiency in field plantings of pecan [Caryaillinoinensis (Wangenh.) K. Koch] (Wood et al., 2004) has led to efforts to identify appropriate management approaches to correct tree deficiency and to identify the causes for Ni deficiency. Evaluation of several inorganic and organic forms of Ni have indicated that solutions from all sources function well to correct deficiencies when timely applied as a foliar spray to affected trees at Ni concentrations >10 mg·L-1. Addition of urea, ammonium nitrate, or nicotinic acid to Ni spray solutions increased apparent foliar uptake from Ni sprays. The lower critical level of Ni, based on foliar analysis, appears to be in the 3-5 mg·L-1 dw range, with the upper critical level appearing to be >50 mg·L-1 dw. The cause of Ni deficiency in soils possessing plenty of Ni is associated with excessive amounts of one or more metals (e.g., Ca, Mg, Fr, Mn, Cu, and Zn) that inhibit Ni uptake and/or utilization. Root damage by nematode feeding and cool/dry soils during early spring also contributes to Ni deficiency. Foliar application of Ni to foliage in the autumn and subsequent appearance of Ni in dormant season shoot tissues indicates that Ni can be mobilized from senescing foliage to dormant season shoots and is therefore available for early spring growth. Evidence indicates that pecan has a higher Ni requirement than most other crop species because it transports nitrogenous substances as ureides. Thus, there is evidence that Ni-metalloenzymes are playing either a direct or indirect role in ureide and nitrogen metabolism. It is postulated that crop species that are most likely to exhibit field level Ni deficiencies are those that transport N as ureides. Candidate crops will be discussed.

Robotic machine learning of anaphora

Our contribution tackles the problem of learning to understand anaphoric references in the context of robotic machine learning; e.g. Get the large screw. Put it in the left hole. Our solution assumes the probabilistic theory of learning spelt out in earlier publications. Associations are formed probabilistically between constituents of the verbal command and constituents of a presupposed internal representation. The theory is extended, as a first step, to anaphora by learning how to distinguish between incorrect surface depth and the correct tree-structure depth of the anaphoric references.

Genetic Distances and Reconstruction of Phylogenetic Trees From Microsatellite DNA

Recently many investigators have used microsatellite DNA loci for studying the evolutionary relationships of closely related populations or species, and some authors proposed new genetic distance measures for this purpose. However, the efficiencies of these distance measures in obtaining the correct tree topology remains unclear. We therefore investigated the probability of obtaining the correct topology (PC) for these new distances as well as traditional distance measures by using computer simulation. We used both the infinite-allele model (IAM) and the stepwise mutation model (SMM), which seem to be appropriate for classical markers and microsatellite loci, respectively. The results show that in both the IAM and SMM Cavalli-Sforza and Edwards' chord distance (Dc) and NEI et al.'s DA distance generally show higher PC values than other distance measures, whether the bottleneck effect exists or not. For estimating evolutionary times, however, NEI'S standard distance and Goldstein et al's (δμ) are more appropriate than other distances. Microsatellite DNA seems to be very useful for clarifying the evolutionary relationships of closely related populations.

TREE FORM: ARCHITECTURAL MODELS DO NOT SUFFICE

Tree forms have been classified according to the aspects of their development that follow relatively invariable programs. However, a prominent variability or plasticity is apparent in the duration and rate of the development of individual shoot apices and the probability that a branch, once formed, remains on the tree. Even though overall form may be predictable, these variable processes have a major role in the detailed structure of tree canopies. It follows that there must be mechanisms that constrain or limit chance events after, rather than before, they have been expressed in tree development. Constraining mechanisms include the inhibition of the development of a branch by another branch that had already started growing. This inhibition is due to both internal signals and to mutual shading. It results in any available space in a canopy being occupied by one or more branches, even though the precise location of these branches is not specified. The development of leading branches is also constrained by a progressive differentiation and growth limitation which is a function of the distance of the apices of these branches from the roots. At both the cellular and the organ levels unpredictable details are a characteristic of advanced, late evolutionary groups. Their prevalence means that architectural models stress an important, but partial, picture. There could be no ideal models nor a correct tree classification; there is no alternative to comparing tree forms in relation to major characteristics chosen according to the problem at hand.