Similarity by compression

We present a simple and effective method for similarity searching in virtual high-throughput screening, requiring only a string-based representation of the molecules (e.g., SMILES) and standard compression software, available on all modern desktop computers. This method utilizes the normalized compression distance, an approximation of the normalized information distance, based on the concept of Kolmogorov complexity. On representative data sets, we demonstrate that compression-based similarity searching can outperform standard similarity searching protocols, exemplified by the Tanimoto coefficient combined with a binary fingerprint representation and data fusion. Software to carry out compression-based similarity is available from our Web site at http://comp.chem.nottingham.ac.uk/download/zippity.     

Circular and linear dichroism of proteins

Circular dichroism (CD) is an important technique in the structural characterisation of proteins, and especially for secondary structure determination. The CD of proteins can be calculated from first principles using the so-called matrix method, with an accuracy which is almost quantitative for helical proteins. Thus, for proteins of unknown structure, CD calculations and experimental data can be used in conjunction to aid structure analysis. Linear dichroism (LD) can be calculated using analogous methodology and has been used to establish the relative orientations of subunits in proteins and protein orientation in an environment such as a membrane. However, simple analysis of LD data is not possible, due to overlapping transitions. So coupling the calculations and experiment is an important strategy. In this paper, the use of LD for the determination of protein orientation and how these data can be interpreted with the aid of calculations, are discussed. We review methods for the calculation of CD spectra, focusing on semiempirical and ab initio parameter sets used in the matrix method. Lastly, a new web interface for online CD and LD calculation is presented.     

Gasanalyse

Ohne Zusammenfassung     

Supervised machine learning algorithms for protein structure classification

We explore automation of protein structural classification using supervised machine learning methods on a set of 11,360 pairs of protein domains (up to 35% sequence identity) consisting of three secondary structure elements. Fifteen algorithms from five categories of supervised algorithms are evaluated for their ability to learn for a pair of protein domains, the deepest common structural level within the SCOP hierarchy, given a one-dimensional representation of the domain structures. This representation encapsulates evolutionary information in terms of sequence identity and structural information characterising the secondary structure elements and lengths of the respective domains. The evaluation is performed in two steps, first selecting the best performing base learners and subsequently evaluating boosted and bagged meta learners. The boosted random forest, a collection of decision trees, is found to be the most accurate, with a cross-validated accuracy of 97.0% and F-measures of 0.97, 0.85, 0.93 and 0.98 for classification of proteins to the Class, Fold, Super-Family and Family levels in the SCOP hierarchy. The meta learning regime, especially boosting, improved performance by more accurately classifying the instances from less populated classes.     

Classroom note: Sawtooth functions

Using MAPLE enables students to consider many examples which would be very tedious to work out by hand. This applies to graph plotting as well as to algebraic manipulation. The challenge is to use these observations to develop the students’ understanding of mathematical concepts. In this note an interesting relationship arising from inverse trigonometric functions is analysed. To understand what is going on students have to develop an understanding of how to deal with inverses where a function is not 1–1, by restricting the domain. The piece of work developed here also provides some interesting exercises in proof by induction.     

Ramsey’s theorem for trees: the polarized tree theorem and notions of stability

We formulate a polarized version of Ramsey’s theorem for trees. For those exponents greater than 2, both the reverse mathematics and the computability theory associated with this theorem parallel that of its linear analog. For pairs, the situation is more complex. In particular, there are many reasonable notions of stability in the tree setting, complicating the analysis of the related results.     

High-voltage microwave-triggered switches

The operation of three types of experimental high voltage microwave-triggered switches developed by the University of St. Andrews and DRA Malvern is described. The POLOTRON and the microwave-triggered three-gap-switch (MTGS) are closing switches and the microwave-triggered tacitron, (MITTON) is a closing and opening switch. The switches are robust hydrogen thyratron type devices, with an annular geometry, where closure is achieved using microwave fields to ionize the low pressure gas within the switch. In the case of the MITTON, the switch is opened by the application of a negative bias voltage to a large area metal grid. The POLOTRON is designed for fast switching applications. Its annular geometry results in an inherent low inductance and anode fall-times of less than 3 ns have been measured for a cold-cathode POLOTRON at charging voltages up to 30 kV. The MTGS, also a cold cathode switch, was developed for use in the DUOTRON transient generator (a voltage doubler). Output voltages of up to 16 kV have been measured at repetition rates of 90 Hz. The MITTON has closing characteristics similar to the POLOTRON and anode voltage fall-times of less than 10 ns have been measured at voltages up to 16 kV. Opening times of 0.5 μs have been achieved with an anode voltage of 10 kV and current of 10 A