Dalius Navakauskas

Proteomic Analysis of Stromal Cells Derived from the Dental Pulp of Human Exfoliated Deciduous Teeth

Human dental pulp derived from exfoliated deciduous teeth has been described as a promising alternative source of multipotent stem cells. While these cells share certain similarities with mesenchymal stem-like cells (MSC) isolated from other tissues, basically they are still poorly characterized. In this study, for the first time, a proteomic map of abundantly expressed proteins in stromal cells derived from the dental pulp of human exfoliated deciduous teeth (SHED) was established. We also analyzed proteomic signatures of 2 clonal strains derived from SHEDs by single-cell cloning. The SHEDs were established from enzyme-disaggregated deciduous dental pulp from 6-year-old children. They had typical fibroblastoid morphology and high colony-forming efficiency index (16.4%). Cloning was performed at the second passage using limiting dilution in a 96-well plate (0.3 cell/well). Differentiation assessment revealed strong osteogenic but no adipogenic potential of the SHEDs in either clonal strain. The cells expressed characteristic antigens of MSC-like cells, including CD73, CD90, CD105, CD146, and did not express hematopoietic markers CD14, CD34, and CD45, as assessed with FACS analysis. For proteomic studies, cytosolic and nuclear proteins were analyzed with 2-dimensional gel electrophoresis (2-DE) and identified using matrix-assisted laser desorption/ionization (MALDI)-time of fl ight (TOF)-mass spectrometry (MS). All proteins were identified with high level of confidence (the lowest sequence coverage was 27%). Identification of highly expressed proteins in SHEDs revealed proteomic profiles very similar to that of MSC-like cells derived from other tissues. We also found a high degree of similarity between proteomic signatures of primary SHEDs and clonal cell strains. Thus, our data confirm a close resemblance between SHEDs and MSC-like cells from other tissues and may serve as starting point for creating-comprehensive proteomic maps.

Feature selection for segmentation of 2-D electrophoresis gel images

Two-dimensional gel electrophoresis is the powerful technique used by biochemists to resolve and visualize protein samples.Commonly gels produced from several samples are analyzed in order to detect changes of protein expression. Thus computer-aided gel image analysis for protein spot detection became the main step in the whole process.Nevertheless accurate automatic spot detection is still difficult due to large variations in spot shape, image background and various inevitable artifacts. In this paper we investigate features of two-dimensional electrophoresis gel images.We look for those image features that will yield good results of protein spot detection done by the Feedforward Multilayer Neural Network. Feature comparison and spot segmentation results are presented and indicate that rotational symmetry features empowers segmentation of saturated and overlapped protein spots.

FPGA based fast Lithuanian isolated word recognition system

The article presents the Lithuanian isolated word recognition system implementation in a FPGA hard-core. The pursued objective is the acceleration of the previous soft-core implementation at both key stages: feature extraction and recognition. The 12-th order cepstral analysis is used to extract speech signal features, while for isolated word recognition a dynamic time warping is used. Implementation completely done in the VHDL hard-core allowed us to 320 times speed-up the signal cepstrum calculation and 348 times - one dynamic time warping comparison with border constraints. The recognition system works in real time and is built on medium class FPGA, operating at 50 MHz main clock frequency. It is tested on 6 times repeated 100 Lithuanian words dictionary. Speaker dependent recognition tests done for 10 speakers yield the 97.7 % average recognition accuracy (with 4.9 % recognition improvement over the previous implementation).

Alpha-Dystrobrevin and its associated proteins in human promyelocytic leukemia cells induced to apoptosis

Dystrobrevin is a dystrophin-related component of the dystrophin-associated protein complex (DAPC). Using alpha-dystrobrevin as indicator, we aimed to elucidate the interaction network of the DAPC with other proteins during apoptosis of promyelocytic HL-60 cells. The precise role(s) of DBs are not known, but we and others have shown that they play a role in intracellular signal transduction and cellular organization. Apoptosis was induced with etoposide in the absence or presence of Z-VAD to block caspase activity, and we then followed the cellular distribution of α-DB and its association with other proteins, using confocal imaging and cell fractions analyses after immune-precipitation with anti-α-DB and mass spectrometry. Confocal imaging revealed distinct spatial relocalizations of α-DB between the cell membrane, cytosol and nucleus after induction of apoptosis. The expression levels of the identified proteins were evaluated with computer-assisted image analysis of the gels. We thus identified associations with structural and transport proteins (tropomyosin, myosin), membrane (ADAM21, syntrophin), ER-Golgi (TGN51, eIF38) and nuclear (Lamins, ribonucleoprotein C1/C2) proteins. These results suggest that apoptosis-induction in HL-60 cells involves not only classical markers of apoptosis but also a network α-DB-associated proteins at the cell membrane, the cytoplasm and nucleus, affecting key cellular transport processes and cellular structure.

A reduced size lattice-ladder neural network

An idea to incorporate filters in the common structure of an artificial neural network (ANN) is motivated by the need to process time-varying signals using ANN. IIR filters are more advantageous than FIR filters - they require less coefficients, can handle long duration signals, however their stability is not guaranteed and requires special attention. We propose new ANN structure, which is based on the lattice-ladder realization of IIR filters incorporated as multilayer perceptron synapses. It is of small size, has simple monitoring of filter stability, trains rapidly, and is easy expandable in filter orders.

The lattice-ladder neuron and its training circuit implementation in FPGA

FPGA implementation of a lattice-ladder multilayer perceptron structure together with its training algorithm in a full scale seems attractive, however there is a lack of preliminary results on the choice of implementation architecture. The aim of this investigation was to get insights on the selected neuron model fixed-point architecture (necessary to use word length) and its complexity (required number of LUT and DSP slices and BRAM size) by the evaluation of the reproduced by lattice-ladder neuron accuracy of bandwidth and central frequency as also as output signal normalized mean error. Thus the second order fixed-point normalized lattice-ladder neuron with its training algorithm was implemented in Artix-7 FPGA. The experiments were performed using various bandwidths and word length constrains. In general increase of word length yielded smaller mean errors. However the limited size BRAM used for trigonometric function LUTs was a bottleneck to improve the precision while doubling the number of DSP slices.

Acceleration of feature extraction for FPGA-based speech recognition

The paper describes a field programmable gate array implementation of the main part of speech recognition system - feature extraction. In order to accelerate recognition the whole cepstral analysis scheme is implemented in hardware by the use of intellectual property cores. Two field programmable gate array devices are used for evaluation. Comparative experimental results of four different implementations are presented. They grounds achieved 29 times faster speech analysis in comparison with software based analysis subsystem.

Histone modifications patterns in tissues and tumours from acute promyelocytic leukemia xenograft model in response to combined epigenetic therapy.

Xenograft models are suitable for in vivo study of leukemias pathogenesis and the preclinical development of anti-leukemia agents but understanding of epigenetic regulatory mechanisms linking to adult cell functions in pathological conditions during different in vivo treatments is yet unknown. In this study, for the first time epigenetic chromatin modifications were characterized in tissues and tumours from murine xenograft model generated using the human acute promyelocytic leukemia (APL) NB4 cells engrafted in immunodeficient NOG mice. Xenografts were subjected to combined epigenetic treatment by histone deacetylase inhibitor Belinostat, histone methyltransferase inhibitor 3-DZNeaplanocin A and all-trans-retinoic acid based on in vitro model, where such combination inhibited NB4 cell growth and enhanced retinoic acid-induced differentiation to granulocytes. Xenotransplantation was assessed by peripheral blood cells counts, the analysis of cell surface markers (CD15, CD33, CD45) and the expression of certain genes (PML-RAR alpha, CSF3, G-CSFR, WT1). The combined treatment prolonged APL xenograft mice survival and prevented tumour formation. The analysis of the expression of histone marks such as acetylation of H4, trimethylation of H3K4, H3K9 and H3K27 in APL xenograft mice tumours and tissues demonstrated tissue-specific changes in the level of histone modifications and the APL prognostic mark, WT1 protein. In summary, the effects of epigenetic agents used in this study were positive for leukemia prevention and linked to a modulation of the chromatin epigenetic environment in adult tissues of malignant organism.

Word Recognition Acceleration by Double Random Seed Matching in Perceptual Cepstrum Error Space

Paper presents an algorithm for acceleration of the dynamic time warping (DTW) based isolated word recognition algorithm. The number of matching operations directly depends on the size of vocabulary. A set of perceptual cepstrum features is calculated for each word and stored in the vocabulary as a reference. Additionally all words (references) are compared between each other using DTW in order to get the reference-to-reference matches. The acceleration of pattern matching is acquired by adaptive search of the pattern reference according to the previous matching results ant reference-to-reference matches. A modified word selection scenario applied for the vocabulary reduces the number of matching operations by 62-70 % in average. The reduction of matching operations allows to use DTW based speech recognition methods in real-time control applications and only need additional 13 % of vocabulary storage space.

Restoration of voiced speech signals preserving prosodic features

The paper deals with the restoration of voiced speech signals that contain clicks, noise, and gaps of sufficient length so that entire phonemes are lost. A particular focus is on restoring the proper prosodic elements: accent and stress. The importance of the problem is grounded by the fact that the meaning of some words (known as homographs) is solely dependent on the prosody. A new restoration method, exploiting the developed simple polynomial accent model with averaged speech signal characteristics of intensity and a fundamental frequency period as its parameters, is proposed. Feasibility of the method is confirmed by experimental investigations of the restoration of both one period and multiple periods of a voiced speech signal, examination of an instantaneous error, the total mean-square-error, and the influence of sampling frequency on the restoration quality.