A probabilistic framework for landmark detection based on phonetic features for automatic speech recognition

A probabilistic framework for a landmark-based approach to speech recognition is presented for obtaining multiple landmark sequences in continuous speech. The landmark detection module uses as input acoustic parameters (APs) that capture the acoustic correlates of some of the manner-based phonetic features. The landmarks include stop bursts, vowel onsets, syllabic peaks and dips, fricative onsets and offsets, and sonorant consonant onsets and offsets. Binary classifiers of the manner phonetic features-syllabic, sonorant and continuant-are used for probabilistic detection of these landmarks. The probabilistic framework exploits two properties of the acoustic cues of phonetic features-(1) sufficiency of acoustic cues of a phonetic feature for a probabilistic decision on that feature and (2) invariance of the acoustic cues of a phonetic feature with respect to other phonetic features. Probabilistic landmark sequences are constrained using manner class pronunciation models for isolated word recognition with known vocabulary. The performance of the system is compared with (1) the same probabilistic system but with mel-frequency cepstral coefficients (MFCCs), (2) a hidden Markov model (HMM) based system using APs and (3) a HMM based system using MFCCs.     

On the hamiltonicity of a planar graph and its vertex-deleted subgraphs

Tutte proved that every planar 4-connected graph is hamiltonian. Thomassen showed that the same conclusion holds for the superclass of planar graphs with minimum degree at least 4 in which all vertex-deleted subgraphs are hamiltonian. We here prove that if in a planar $n$-vertex graph with minimum degree at least 4 at least $n-5$ vertex-deleted subgraphs are hamiltonian, then the graph contains two hamiltonian cycles, but that for every there exists a nonhamiltonian polyhedral $n$-vertex graph with minimum degree at least 4 containing $cn$ hamiltonian vertex-deleted subgraphs. Furthermore, we study the hamiltonicity of planar triangulations and their vertex-deleted subgraphs as well as Bondy's meta-conjecture, and prove that a polyhedral graph with minimum degree at least 4 in which all vertex-deleted subgraphs are traceable, must itself be traceable.     

Side-Chain Chemistry Governs Hierarchical Order of Charge-Complementary β-sheet Peptide Coassemblies

Self-assembly of proteinaceous biomolecules into functional materials with ordered structures that span length scales is common in nature yet remains a challenge with designer peptides under ambient conditions. This report demonstrates how charged side-chain chemistry affects the hierarchical co-assembly of a family of charge-complementary β-sheet-forming peptide pairs known as CATCH(X+/Y−) at physiologic pH and ionic strength in water. In a concentration-dependent manner, the CATCH(6K+) (Ac-KQKFKFKFKQK-Am) and CATCH(6D−) (Ac-DQDFDFDFDQD-Am) pair formed either β-sheet-rich microspheres or β-sheet-rich gels with a micron-scale plate-like morphology, which were not observed with other CATCH(X+/Y−) pairs. This hierarchical order was disrupted by replacing D with E, which increased fibril twisting. Replacing K with R, or mutating the N- and C-terminal amino acids in CATCH(6K+) and CATCH(6D−) to Qs, increased observed co-assembly kinetics, which also disrupted hierarchical order. Due to the ambient assembly conditions, active CATCH(6K+)-green fluorescent protein fusions could be incorporated into the β-sheet plates and microspheres formed by the CATCH(6K+/6D−) pair, demonstrating the potential to endow functionality.     

Conformational analysis of siloxane-based enzyme-mimic precursors[1]

The conformational flexibility of six hybrid organodisiloxane oligomers were studied using the Low Mode-Monte Carlo conformational search method with the MM2* force field and the Generalized Born/Surface Area continuum solvent model for water. These systems have enzyme-like properties as synthetic acyltransferases and contain aminopyridine groups in various states of protonation. An ensemble of low energy structures was generated and used to investigate the dependence of molecular shape and flexibility on protonation, which plays an important role in catalyst solubility and self-association. The results as measured by the number of unique conformations, end-to-end or longest intramolecular distance and radius of gyration of the conformational point cloud indicate that the number of protonated pyridines plays a significant role in the overall molecular shape. A similar study was also carried out on various POSS-substitutive organodisiloxane oligomers.     

High-Throughput Optimization of Photochemical Reactions using Segmented-Flow Nanoelectrospray-Ionization Mass Spectrometry**

Within the realm of drug discovery, high-throughput experimentation techniques enable the rapid optimization of reactions and expedited generation of drug compound libraries for biological and pharmacokinetic evaluation. Herein we report the development of a segmented flow mass spectrometry-based platform to enable the rapid exploration of photoredox reactions for early-stage drug discovery. Specifically, microwell plate-based photochemical reaction screens were reformatted to segmented flow format to enable delivery to nanoelectrospray ionization-mass spectrometry analysis. This approach was demonstrated for the late-stage modification of complex drug scaffolds, as well as the subsequent structure–activity relationship evaluation of synthesized analogs. This technology is anticipated to expand the robust capabilities of photoredox catalysis in drug discovery by enabling high-throughput library diversification.