TBMaLT, a flexible toolkit for combining tight-binding and machine learning.

Tight-binding approaches, especially the Density Functional Tight-Binding (DFTB) and the extended tight-binding schemes, allow for efficient quantum mechanical simulations of large systems and long-time scales. They are derived from ab initio density functional theory using pragmatic approximations and some empirical terms, ensuring a fine balance between speed and accuracy. Their accuracy can be improved by tuning the empirical parameters using machine learning techniques, especially when information about the local environment of the atoms is incorporated. As the significant quantum mechanical contributions are still provided by the tight-binding models, and only short-ranged corrections are fitted, the learning procedure is typically shorter and more transferable as it were with predicting the quantum mechanical properties directly with machine learning without an underlying physically motivated model. As a further advantage, derived quantum mechanical quantities can be calculated based on the tight-binding model without the need for additional learning. We have developed the open-source framework-Tight-Binding Machine Learning Toolkit-which allows the easy implementation of such combined approaches. The toolkit currently contains layers for the DFTB method and an interface to the GFN1-xTB Hamiltonian, but due to its modular structure and its well-defined interfaces, additional atom-based schemes can be implemented easily. We are discussing the general structure of the framework, some essential implementation details, and several proof-of-concept applications demonstrating the perspectives of the combined methods and the functionality of the toolkit.

Simultaneous on-line vacuum single- and multi-photon ionization on an orthogonal acceleration time-of-flight mass spectrometer platform.

The development of orthogonal acceleration time-of-flight (oa-tof) technology is driven forward due to higher mass accuracy and resolving power than conventional linear/reflectron tof instruments. This is achieved with a more accurate definition of starting energies and coordinates of ions by spatial separation of ion generation and orthogonal ion extraction. Consequently, the ability to cover the whole mass spectral range without scanning is not given anymore. Therefore, continuous ion sources are favored for ensuring high duty cycles and thus high temporal resolution. For pulsed ion sources, high repetition rates are mandatory for covering large m/z ranges without losing their high temporal resolution. We have combined an oa-tof with deuterium lamp single-photon ionization (SPI) as a continuous ion source together with a pulsed 2000 Hz excimer (KrF) laser for resonance enhanced multi-photon ionization (REMPI). These two ionization techniques can be used simultaneously. To the best of our knowledge, this system is the first of its kind in combining a vacuum pulsed ionization source with an oa-tof instrument without any other ion storage hardware. The combination of a soft broadband ionization for organics (SPI) in combination with a very sensitive and selective soft ionization (REMPI) can be used for covering the whole mass range or in targeted on-line monitoring cases one or several smaller mass ranges. To demonstrate the simultaneous SPI/REMPI-oa-tof technique, two applicative areas are explored: on-line monitoring of coffee roast gas emissions and e-cigarette vapor. The complementary information from SPI and REMPI signals are combined in a way to exploit the advantages of both ionization types. In a further development step, a second data acquisition card is built into the system. This modification allows the independent storage of data from both ionization methods without mixing. For demonstration, a third example with a GC measurement is provided. The last example shows the possibility of modified sensitivities for different mass regions in REMPI data acquisition without affecting the SPI channel. The newly developed system shows high robustness in terms of measurements in real industrial environments. The simultaneous measurement technique provides a higher density of information in a single measurement, saving time and resources.

DFTB+, a software package for efficient approximate density functional theory based atomistic simulations.

DFTB+ is a versatile community developed open source software package offering fast and efficient methods for carrying out atomistic quantum mechanical simulations. By implementing various methods approximating density functional theory (DFT), such as the density functional based tight binding (DFTB) and the extended tight binding method, it enables simulations of large systems and long timescales with reasonable accuracy while being considerably faster for typical simulations than the respective ab initio methods. Based on the DFTB framework, it additionally offers approximated versions of various DFT extensions including hybrid functionals, time dependent formalism for treating excited systems, electron transport using non-equilibrium Green's functions, and many more. DFTB+ can be used as a user-friendly standalone application in addition to being embedded into other software packages as a library or acting as a calculation-server accessed by socket communication. We give an overview of the recently developed capabilities of the DFTB+ code, demonstrating with a few use case examples, discuss the strengths and weaknesses of the various features, and also discuss on-going developments and possible future perspectives.

High-resolution time and spatial imaging of tobacco and its pyrolysis products during a cigarette puff by microprobe sampling photoionisation mass spectrometry.

The time- and space-resolved chemical signatures of gases and vapours formed in solid-state combustion processes are difficult to examine using recent analytical techniques. A machine-smoked cigarette represents a very reproducible model system for dynamic solid-state combustion. By using a special sampling system (microprobe unit) that extracts the formed gases from inside of the burning cigarette, which is coupled to a photoionisation mass spectrometer, it was possible to study the evolution of organic gases during a 2-s cigarette puff. The concentrations of various pyrolysis and combustion products such as 1,3-butadiene, toluene, acetaldehyde and phenol were monitored on-line at different sampling points within cigarettes. A near-microscopic-scale spatial resolution and a 200-ms time resolution were achieved. Finally, the recorded information was combined to generate time-resolved concentration maps, showing the formation and destruction zones of the investigated compounds in the burning cigarette. The combustion zone at the tip of cigarette, where e.g. 1,3-butadiene is predominately formed, was clearly separable from the pyrolysis zones. Depending on the stability of the precursor (e.g. lignin or cellulose), the position of pyrolytic formation varies. In conclusion, it was demonstrated that soft photoionisation mass spectrometry in conjunction with a microprobe sampling device can be used for time- and space-resolved analysis of combustion and pyrolysis reactions. In addition to studies on the model cigarette, further model systems may be studied with this approach. This may include further studies on the combustion of biomass or coal chunks, on heterogeneously catalysed reactions or on spray, dust and gas combustion processes.

Rapid on-site detection of explosives on surfaces by ambient pressure laser desorption and direct inlet single photon ionization or chemical ionization mass spectrometry.

Considering current security issues, powerful tools for detection of security-relevant substances such as traces of explosives and drugs/drug precursors related to clandestine laboratories are required. Especially in the field of detection of explosives and improvised explosive devices, several relevant compounds exhibit a very low vapor pressure. Ambient pressure laser desorption is proposed to make these substances available in the gas phase for the detection by adapted mass spectrometers or in the future with ion-mobility spectrometry as well. In contrast to the state-of-the-art thermal desorption approach, by which the sample surface is probed for explosive traces by a wipe pad being transferred to a thermal desorber unit, by the ambient pressure laser desorption approach presented here, the sample is directly shockwave ablated from the surface. The laser-dispersed molecules are sampled by a heated sniffing capillary located in the vicinity of the ablation spot into the mass analyzer. This approach has the advantage that the target molecules are dispersed more gently than in a thermal desorber unit where the analyte molecules may be decomposed by the thermal intake. In the technical realization, the sampling capillary as well as the laser desorption optics are integrated in the tip of an endoscopic probe or a handheld sampling module. Laboratory as well as field test scenarios were performed, partially in cooperation with the Federal Criminal Police Office (Bundeskriminalamt, BKA, Wiesbaden, Germany), in order to demonstrate the applicability for various explosives, drugs, and drug precursors. In this work, we concentrate on the detection of explosives. A wide range of samples and matrices have been investigated successfully.

The three-dimensional structure of cystathionine beta-lyase from Arabidopsis and its substrate specificity.

The pyridoxal 5'-phosphate-dependent enzyme cystathionine beta-lyase (CBL) catalyzes the penultimate step in the de novo biosynthesis of Met in microbes and plants. Absence of CBL in higher organisms makes it an important target for the development of antibiotics and herbicides. The three-dimensional structure of cystathionine beta-lyase from Arabidopsis was determined by Patterson search techniques, using the structure of tobacco (Nicotiana tabacum) cystathionine gamma-synthase as starting point. At a resolution of 2.3 A, the model was refined to a final crystallographic R-factor of 24.9%. The overall structure is very similar to other pyridoxal 5'-phosphate-dependent enzymes of the gamma-family. Exchange of a few critical residues within the active site causes the different substrate preferences between Escherichia coli and Arabidopsis CBL. Loss of interactions at the alpha-carboxyl site is the reason for the poorer substrate binding of Arabidopsis CBL. In addition, the binding pocket of Arabidopsis CBL is larger than that of E. coli CBL, explaining the similar binding of L-cystathionine and L-djenkolate in Arabidopsis CBL in contrast to E. coli CBL, where the substrate binding site is optimized for the natural substrate cystathionine.

Structural investigations of the highly flexible recombinant ribosomal protein L12 from Thermotoga maritima.

Ribosomal protein L7/L12, the only multicopy component of the ribosome, is involved in translation factor binding and in the ribosomal GTPase center. The gene for L7/L12 from Thermotoga maritima was cloned and the protein expressed at high levels in Escherichia coli. Purification of L7/L12 was achieved under non-denaturing conditions via heat treatment and two chromatographic steps. Circular dichroism melting profiles were monitored at 222 nm, showing the melting temperature of the protein at pH 7.5 around 110 degrees C, compared to approximately 60 degrees C for the highly homologous Escherichia coli protein. The unfolding was reversible and renaturation closely followed the path of the thermal melting. Dynamic light scattering, gel filtration chromatography, and crosslinking experiments suggested that under physiological buffer conditions Thermotoga maritima L7/L12 exists as a tetramer. The protein was crystallized under two conditions, yielding an orthorhombic (C222(1)) and a cubic (12(1)3) space group with an estimated two and three to four L7/L12 molecules per asymmetric unit, respectively. The crystals contained the full-length protein, and cryogenic buffers were developed which improved the mosaic spreads and the resolution limits. For the structure solution isoleucine was mutated to methionine at two separate positions, the mutant forms expressed as selenomethionine variants and crystallized.

Characterization of recombinant Arabidopsis thaliana threonine synthase.

Threonine synthase (TS) catalyses the last step in the biosynthesis of threonine, the pyridoxal 5'-phosphate dependent conversion of L-homoserine phosphate (HSerP) into L-threonine and inorganic phosphate. Recombinant Arabidopsis thaliana TS (aTS) was characterized to compare a higher plant TS with its counterparts from Escherichia coli and yeast. This comparison revealed several unique properties of aTS: (a) aTS is a regulatory enzyme whose activity was increased up to 85-fold by S-adenosyl-L-methionine (SAM) and specifically inhibited by AMP; (b) HSerP analogues shown previously to be potent inhibitors of E. coli TS failed to inhibit aTS; and (c) aTS was a dimer, while the E. coli and yeast enzymes are monomers. The N-terminal region of aTS is essential for its regulatory properties and protects against inhibition by HSerP analogues, as an aTS devoid of 77 N-terminal residues was neither activated by SAM nor inhibited by AMP, but was inhibited by HSerP analogues. The C-terminal region of aTS seems to be involved in dimer formation, as the N-terminally truncated aTS was also found to be a dimer. These conclusions are supported by a multiple amino-acid sequence alignment, which revealed the existence of two TS subfamilies. aTS was classified as a member of subfamily 1 and its N-terminus is at least 35 residues longer than those of any nonplant TS. Monomeric E. coli and yeast TS are members of subfamily 2, characterized by C-termini extending about 50 residues over those of subfamily 1 members. As a first step towards a better understanding of the properties of aTS, the enzyme was crystallized by the sitting drop vapour diffusion method. The crystals diffracted to beyond 0.28 nm resolution and belonged to the space group P222 (unit cell parameters: a = 6.16 nm, b = 10.54 nm, c = 14.63 nm, alpha = beta = gamma = 90 degrees).

Isolation, cloning, sequence analysis and X-ray structure of dimethyl sulfoxide/trimethylamine N-oxide reductase from Rhodobacter capsulatus.

The periplasmic enzyme dimethyl sulfoxide/trimethylamine N-oxide reductase (DMSOR/TMAOR) from the photosynthetic purple bacterium Rhodobacter capsulatus functions as the terminal electron acceptor in its respiratory chain. The enzyme catalyzes the reduction of highly oxidized substrates like dimethyl sulfoxide (DMSO) or trimethylamine N-oxide (TMAO). At a molybdenum redox centre, two single electrons are transferred from cytochrome c556 to the substrate, e.g. DMSO, generating dimethyl sulfide (DMS) and water. The operon encoding this enzyme was isolated, cloned and sequenced, and its chromosomal location determined. It was shown by analytical and crystallographic data that DMSOR and TMAOR are identical enzymes. Degenerate primers were derived from short peptide sequences and a 700 bp fragment was amplified by nested PCR, subsequently cloned and radioactively labeled to screen a prepared lambda DASH library. Positive lambda clones were subcloned into pBluescript and subsequently transformed into Escherichia coli to sequence the DMSOR/TMAOR operon. By an optimized protein purification high yields (5 mg protein/l culture) with a specific activity of 30 U/mg were obtained. The molecular mass was experimentally determined by electrospray mass spectroscopy (MS) to be 85034 Da and from the deduced amino acid sequence of the apoenzyme to be 85033 Da. The enzyme was crystallized in space group P4(1)2(1)2 with unit cell dimensions of a = b = 80.7 A and c = 229.2 A diffracting beyond 1.8 A. The three-dimensional structure was solved by a combination of multiple isomorphous replacement (MIR) and molecular replacement techniques. The atomic model was refined to an R-factor of 0.169 for 57394 independent reflections. The spherical protein consists of four domains with a funnel-like cavity that leads to the freely accessible metal-ion redox center. The sole bis(molybdopterin guanine dinucleotide)molybdenum cofactor (1541 Da) of the single chain protein has the molybdenum ion bound to the cis-dithiolene group of only one molybdopterin guanine dinucleotide (MGD) molecule. In addition, two oxo ligands and the oxygen of a serine side chain are bound to the molybdenum ion.

Isolation, cloning, sequence analysis and localization of the operon encoding dimethyl sulfoxide/trimethylamine N-oxide reductase from Rhodobacter capsulatus.

The operon encoding the periplasmic enzymes dimethyl sulfoxide reductase (DMSOR) and trimethylamine N-oxide reductase (TMAOR) from the purple, non-sulfur, photosynthetic bacterium Rhodobacter capsulatus was isolated, cloned and sequenced, and its chromosomal location determined. It was shown by analytical and crystallographic data that DMSOR and TMAOR are identical enzymes. Degenerate primers were derived from short peptide sequences generated by automated Edman degradation and a 700 bp fragment was amplified by nested PCR, subsequently cloned and radioactively labeled to screen a prepared lambda DASH library. Positive lambda clones were subcloned into pBluescript and subsequently transformed into Escherichia coli to sequence the DMSOR/ TMAOR operon. The promoter consisted of an A + T-rich region followed by a -35 region, a putative ribosome binding site, and a leader sequence of 13 amino acid residues. The transcription terminator was a G + C-rich dyad sequence capable of forming a hairpin structure, which may act rho-independently. An optimized protein purification of the wild-type enzyme is also described, giving high yields (5 mg protein per liter of culture) and a specific activity of 30 units/mg. The molecular mass was determined by electrospray mass spectrometry to be 85,034 Da; from the deduced amino acid sequence the molecular mass of the apoenzyme was 85,033 Da.

The validity generalization of distance run tests.

The validity generalization model was used to examine the generalizability of the validity of distance run tests as measures of cardiorespiratory function. A literature search was conducted to identify studies in which distance run test scores were correlated with VO2 max scores obtained from a stress test on a treadmill. The data base was limited to runs of at least one mile or 9 min and the VO2 max scores expressed as mL.kg-1.min-1. The concurrent validity of distance run tests was not shown to be generalizable across all situations. When the data were analyzed by gender, the validity of distance run tests was more generalizable for women than for men. In the analysis of test scores for boys and girls, the validities appeared to be generalizable, although the results should be interpreted with caution owing to small samples sizes.