Towards electronic health record-based medical knowledge graph construction, completion, and applications: A literature study.

With the growth of data and intelligent technologies, the healthcare sector opened numerous technology that enabled services for patients, clinicians, and researchers. One major hurdle in achieving state-of-the-art results in health informatics is domain-specific terminologies and their semantic complexities. A knowledge graph crafted from medical concepts, events, and relationships acts as a medical semantic network to extract new links and hidden patterns from health data sources. Current medical knowledge graph construction studies are limited to generic techniques and opportunities and focus less on exploiting real-world data sources in knowledge graph construction. A knowledge graph constructed from Electronic Health Records (EHR) data obtains real-world data from healthcare records. It ensures better results in subsequent tasks like knowledge extraction and inference, knowledge graph completion, and medical knowledge graph applications such as diagnosis predictions, clinical recommendations, and clinical decision support. This review critically analyses existing works on medical knowledge graphs that used EHR data as the data source at (i) representation level, (ii) extraction level (iii) completion level. In this investigation, we found that EHR-based knowledge graph construction involves challenges such as high complexity and dimensionality of data, lack of knowledge fusion, and dynamic update of the knowledge graph. In addition, the study presents possible ways to tackle the challenges identified. Our findings conclude that future research should focus on knowledge graph integration and knowledge graph completion challenges.

Breast cancer stage prediction: a computational approach guided by transcriptome analysis.

Breast cancer is the second leading cancer among women in terms of mortality rate. In recent years, its incidence frequency has been continuously rising across the globe. In this context, the new therapeutic strategies to manage the deadly disease attracts tremendous research focus. However, finding new prognostic predictors to refine the selection of therapy for the various stages of breast cancer is an unattempted issue. Aberrant expression of genes at various stages of cancer progression can be studied to identify specific genes that play a critical role in cancer staging. Moreover, while many schemes for subtype prediction in breast cancer have been explored in the literature, stage-wise classification remains a challenge. These observations motivated the proposed two-phased method: stage-specific gene signature selection and stage classification. In the first phase, meta-analysis of gene expression data is conducted to identify stage-wise biomarkers that were then used in the second phase of cancer classification. From the analysis, 118, 12 and 4 genes respectively in stage I, stage II and stage III are determined as potential biomarkers. Pathway enrichment, gene network and literature analysis validate the significance of the identified genes in breast cancer. In this study, machine learning methods were combined with principal component and posterior probability analysis. Such a scheme offers a unique opportunity to build a meaningful model for predicting breast cancer staging. Among the machine learning models compared, Support Vector Machine (SVM) is found to perform the best for the selected datasets with an accuracy of 92.21% during test data evaluation. Perhaps, biomarker identification performed here for stage-specific cancer treatment would be a meaningful step towards predictive medicine. Significantly, the determination of correct cancer stage using the proposed 134 gene signature set can possibly act as potential target for breast cancer therapeutics.

Cytopathological image analysis using deep-learning networks in microfluidic microscopy.

Cytopathologic testing is one of the most critical steps in the diagnosis of diseases, including cancer. However, the task is laborious and demands skill. Associated high cost and low throughput drew considerable interest in automating the testing process. Several neural network architectures were designed to provide human expertise to machines. In this paper, we explore and propose the feasibility of using deep-learning networks for cytopathologic analysis by performing the classification of three important unlabeled, unstained leukemia cell lines (K562, MOLT, and HL60). The cell images used in the classification are captured using a low-cost, high-throughput cell imaging technique: microfluidics-based imaging flow cytometry. We demonstrate that without any conventional fine segmentation followed by explicit feature extraction, the proposed deep-learning algorithms effectively classify the coarsely localized cell lines. We show that the designed deep belief network as well as the deeply pretrained convolutional neural network outperform the conventionally used decision systems and are important in the medical domain, where the availability of labeled data is limited for training. We hope that our work enables the development of a clinically significant high-throughput microfluidic microscopy-based tool for disease screening/triaging, especially in resource-limited settings.

Optimization of antioxidant efficacy of a deflavored and decolorized rosemary extract: effect of carnosol content on the oxidative stability of paprika colored beef patties.

Considering the significance of natural antioxidants to preserve meat, the present study was undertaken to evaluate the efficacy of a deflavored and decolorised extract of rosemary (StabilRose™) for the production and preservation of naturally colored fresh meat. Oxidative rancidity of meat and color degradation of paprika oleoresin were exploited as model systems and compared with butylated hydroxyanisole (BHA). The results showed similar efficacy for 3% carnosic acid extract and BHA, with further enhancement in efficacy with respect to the carnosic acid content. A synergetic antioxidant effect of carnosol on carnosic acid content was also noticed to an extent of 1:1 (w/w) ratio, and further increase in carnosol content showed no improvement in the antioxidant efficacy. Finally, stabilized paprika and optimized rosemary extract containing carnosic acid and carnosol in 1:1 (w/w) ratio was successfully applied to produce naturally colored meat suitable for storage at 4 ± 1 °C.

Femtosecond supercontinuum generation in water in the vicinity of absorption bands.

We show that it is possible to overcome the perceived limitations caused by absorption bands in water so as to generate supercontinuum (SC) spectra in the anomalous dispersion regime that extend well beyond 2000 nm wavelength. By choosing a pump wavelength within a few hundred nanometers above the zero-dispersion wavelength of 1048 nm, initial spectral broadening extends into the normal dispersion regime and, in turn, the SC process in the visible strongly benefits from phase-matching and matching group velocities between dispersive radiation and light in the anomalous dispersion regime. Some of the SC spectra are shown to encompass two and a half octaves.

Framework for morphometric classification of cells in imaging flow cytometry.

Imaging flow cytometry is an emerging technology that combines the statistical power of flow cytometry with spatial and quantitative morphology of digital microscopy. It allows high-throughput imaging of cells with good spatial resolution, while they are in flow. This paper proposes a general framework for the processing/classification of cells imaged using imaging flow cytometer. Each cell is localized by finding an accurate cell contour. Then, features reflecting cell size, circularity and complexity are extracted for the classification using SVM. Unlike the conventional iterative, semi-automatic segmentation algorithms such as active contour, we propose a noniterative, fully automatic graph-based cell localization. In order to evaluate the performance of the proposed framework, we have successfully classified unstained label-free leukaemia cell-lines MOLT, K562 and HL60 from video streams captured using custom fabricated cost-effective microfluidics-based imaging flow cytometer. The proposed system is a significant development in the direction of building a cost-effective cell analysis platform that would facilitate affordable mass screening camps looking cellular morphology for disease diagnosis.

A Novel Extract of Fenugreek Husk (FenuSMART™) Alleviates Postmenopausal Symptoms and Helps to Establish the Hormonal Balance: A Randomized, Double-Blind, Placebo-Controlled Study.

Despite the widespread use of hormone replacement therapy, various reports on its side effects have generated an increasing interest in the development of safe natural agents for the management of postmenopausal discomforts. The present randomized, double-blinded, placebo-controlled study investigated the effect of 90-day supplementation of a standardized extract of fenugreek (Trigonella foenum-graecum) (FenuSMART™), at a dose of 1000 mg/day, on plasma estrogens and postmenopausal discomforts. Eighty-eight women having moderate to severe postmenopausal discomforts and poor quality of life (as evidenced from the scores of Greene Climacteric Scale, short form SF-36® and structured medical interview) were randomized either to extract-treated (n = 44) or placebo (n = 44) groups. There was a significant (p < 0.01) increase in plasma estradiol (120%) and improvements on various postmenopausal discomforts and quality of life of the participants in the extract-treated group, as compared with the baseline and placebo. While 32% of the subjects in the extract group reported no hot flashes after supplementation, the others had a reduction to one to two times per day from the baseline stages of three to five times a day. Further analysis of haematological and biochemical parameters revealed the safety of the extract and its plausible role in the management of lipid profile among menopausal women. Copyright © 2016 John Wiley & Sons, Ltd.

Jahn-Teller instability in cationic boron and carbon buckyballs B80+ and C60+: a comparative study.

This paper investigates the Jahn-Teller effect in the icosahedral cation B(80)(+) and compares the descent in symmetry with that in C(60)(+). For both cations the icosahedral ground state is a (2)H(u) state, which exhibits a H ⊗ (g ⊕ 2h) Jahn-Teller instability. A detailed construction of the potential energy surface of B(80)(+) using different DFT methods including B3LYP/6-31G(d), VWN/6-31G(d), PBE/TZP and PBE/6-31G(d) shows that, contrary to C(60)(+), which prefers D(5d) symmetry, the ground state of B(80)(+) adopts S(6) point group symmetry. A D(3d) structure is identified as a saddle point among the S(6) minima of B(80)(+). The distortion of D(3d) to S(6) in B(80)(+) is attributed to a superposition of Jahn-Teller and pseudo-Jahn-Teller effects. Imaginary modes, transforming as the g(g) representation, which are present in neutral icosahedral B(80), form the dominant symmetry breaking active modes. The pronounced difference between the JT effects in the boron and carbon buckyball cations is due to the plasticity of the boron caps. The calculated Jahn-Teller stabilization of B(80)(+) is nearly 1549 cm(-1) (PBE/TZP), which exceeds the stabilization of 596 cm(-1) computed for C(60)(+) at the same level.

DeepBindPPI: Protein-Protein Binding Site Prediction Using Attention Based Graph Convolutional Network.

Due to the importance of protein-protein interactions in defence mechanism of living body, attempts were made to investigate its attributes, including, but not limited to, binding affinity, and binding region. Contemporary strategies for binding site prediction largely resort to deep learning techniques but turned out to be low precision models. As laboratory experiments for drug discovery tasks utilize this information, increased false positives devalue the computational methods. This emphasize the need to develop enhanced strategies. DeepBindPPI employs deep learning technique to predict the binding regions of proteins, particularly antigen-antibody interaction sites. The results obtained are applied in a docking environment to confirm their correctness. An integration of graph convolutional network with attention mechanism predicts interacting amino acids with improved precision. The model learns the determining factors in interaction from a general pool of proteins and is then fine-tuned using antigen-antibody data. Comparison of the proposed method with existing techniques shows that the developed model has comparable performance. The use of a separate spatial network clearly improved the precision of the proposed method from 0.4 to 0.5. An attempt to utilize the interface information for docking using the HDOCK server gives promising results, with high-quality structures appearing in the top10 ranks.

Radiation damage by extensive local water ionization from two-step electron-transfer-mediated decay of solvated ions.

Biomolecular radiation damage is largely mediated by radicals and low-energy electrons formed by water ionization rather than by direct ionization of biomolecules. It was speculated that such an extensive, localized water ionization can be caused by ultrafast processes following excitation by core-level ionization of hydrated metal ions. In this model, ions relax via a cascade of local Auger-Meitner and, importantly, non-local charge- and energy-transfer processes involving the water environment. Here, we experimentally and theoretically show that, for solvated paradigmatic intermediate-mass Al3+ ions, electronic relaxation involves two sequential solute-solvent electron transfer-mediated decay processes. The electron transfer-mediated decay steps correspond to sequential relaxation from Al5+ to Al3+ accompanied by formation of four ionized water molecules and two low-energy electrons. Such charge multiplication and the generated highly reactive species are expected to initiate cascades of radical reactions.

Hydroxylation catalysis by mononuclear and dinuclear iron oxo catalysts: a methane monooxygenase model system versus the Fenton reagent Fe(IV)O(H2O)5(2+).

Hydroxylation of aliphatic C-H bonds is a chemically and biologically important reaction, which is catalyzed by the oxidoiron group FeO(2+) in both mononuclear (heme and nonheme) and dinuclear complexes. We investigate the similarities and dissimilarities of the action of the FeO(2+) group in these two configurations, using the Fenton-type reagent [FeO(2+) in a water solution, FeO(H(2)O)(5)(2+)] and a model system for the methane monooxygenase (MMO) enzyme as representatives. The high-valent iron oxo intermediate MMOH(Q) (compound Q) is regarded as the active species in methane oxidation. We show that the electronic structure of compound Q can be understood as a dimer of two Fe(IV)O(2+) units. This implies that the insights from the past years in the oxidative action of this ubiquitous moiety in oxidation catalysis can be applied immediately to MMOH(Q). Electronically the dinuclear system is not fundamentally different from the mononuclear system. However, there is an important difference of MMOH(Q) from FeO(H(2)O)(5)(2+): the largest contribution to the transition state (TS) barrier in the case of MMOH(Q) is not the activation strain (which is in this case the energy for the C-H bond lengthening to the TS value), but it is the steric hindrance of the incoming CH(4) with the ligands representing glutamate residues. The importance of the steric factor in the dinuclear system suggests that it may be exploited, through variation in the ligand framework, to build a synthetic oxidation catalyst with the desired selectivity for the methane substrate.

Ligand Based Virtual Screening Using Self-organizing Maps.

Conventional drug discovery methods rely primarily on in-vitro experiments with a target molecule and an extensive set of small molecules to choose the suitable ligand. The exploration space for the selected ligand being huge; this approach is highly time-consuming and requires high capital for facilitation. Virtual screening, a computational technique used to reduce this search space and identify lead molecules, can speed up the drug discovery process. This paper proposes a ligand-based virtual screening method using an artificial neural network called self-organizing map (SOM). The proposed work uses two SOMs to predict the active and inactive molecules separately. This SOM based technique can uniquely label a small molecule as active, inactive, and undefined as well. This can reduce the number of false positives in the screening process and improve recall; compared to support vector machine and random forest based models. Additionally, by exploiting the parallelism present in the learning and classification phases of a SOM, a graphics processing unit (GPU) based model yields much better execution time. The proposed GPU-based SOM tool can successfully evaluate a large number of molecules in training and screening phases. The source code of the implementation and related files are available at https://github.com/jayarajpbalakrishnan/2_SOM_SCREEN.

The leapfrog principle for boron fullerenes: a theoretical study of structure and stability of B112.

Two leapfrog isomers of a B(112) boron fullerene are constructed from small C(28) fullerenes (T(d) and D(2) symmetries) by the leapfrog transformation combined with omnicapping of the new hexagons. Their electronic structure is analyzed using the density functional theory at the B3LYP/SVP and BHLYP/SVP levels. Both isomers are characterized as minima on the potential energy hypersurface with a HOMO-LUMO gap at B3LYP/SVP of 1.7 eV and 1.6 eV (3.1 and 3.0 eV at BHLYP/SVP), respectively. The optimized structure of the helical D(2)-leapfrog is asymmetric, due to radial displacements of the capping atoms. The computed cohesive energies amount to -4.2 eV (∼0.04 eV lower than B(80)). The B(112) isomers are isoelectronic to T(d)-C(84) and D(2)-C(84), and HOMO and LUMO orbitals in both isomers closely resemble those of their C(84) homologues. Energetic stability of leapfrog boron fullerenes depends on the isolation of empty hexagon criterion, which is defined by the empty hexagon index based on the total number of empty hexagon pairs and empty hexagon-pentagon fused pairs. The switch of the cap atom to the nearest or farthest empty hexagon destabilizes the cage by 1.6 and 2.7 eV, respectively. The destabilization becomes more enhanced in non-leapfrog structures wherein more caps are displaced.

An integrated method for identifying essential proteins from multiplex network model of protein-protein interactions.

Cell survival requires the presence of essential proteins. Detection of essential proteins is relevant not only because of the critical biological functions they perform but also the role played by them as a drug target against pathogens. Several computational techniques are in place to identify essential proteins based on protein-protein interaction (PPI) network. Essential protein detection using only physical interaction data of proteins is challenging due to its inherent uncertainty. Hence, in this work, we propose a multiplex network-based framework that incorporates multiple protein interaction data from their physical, coexpression and phylogenetic profiles. An extended version termed as multiplex eigenvector centrality (MEC) is used to identify essential proteins from this network. The methodology integrates the score obtained from the multiplex analysis with subcellular localization and Gene Ontology information and is implemented using Saccharomyces cerevisiae datasets. The proposed method outperformed many recent essential protein prediction techniques in the literature.

ERBB3-induced furin promotes the progression and metastasis of ovarian cancer via the IGF1R/STAT3 signaling axis.

High-grade serous carcinoma, accounts for up to 70% of all ovarian cases. Furin, a proprotein convertase, is highly expressed in high-grade serous carcinoma of ovarian cancer patients, and its expression is even higher in tumor omentum than in normal omentum, the preferred site of ovarian cancer metastasis. The proteolytic actions of this cellular endoprotease help the maturation of several important precursors of protein substrates and its levels increase the risk of several cancer. We show that furin activates the IGF1R/STAT3 signaling axis in ovarian cancer cells. Conversely, furin knockdown downregulated IGF1R-ß and p-STAT3 (Tyr705) expression. Further, silencing furin reduced tumor cell migration and invasion in vitro and tumor growth and metastasis in vivo. Collectively, our findings show that furin can be an effective therapeutic target for ovarian cancer prevention or treatment.

Experimental detection and theoretical characterization of germanium-doped lithium clusters Li(n)Ge (n = 1-7).

We report a combined experimental and quantum chemical study of the small neutral and cationic germanium-doped lithium clusters Li(n)Ge(0,+) (n = 1-7). The clusters were detected by time-of-flight mass spectrometry after laser vaporization and ionization. The molecular geometries and electronic structures of the clusters were investigated using quantum chemical calculations at the DFT/B3LYP and CCSD(T) levels with the aug-cc-pVnZ basis sets. While Li3Ge(0,+) and Li4Ge+ prefer planar structures, the clusters from Li4Ge to Li7Ge and the corresponding cations (except Li4Ge+) exhibit nonplanar forms. Clusters having from 4 to 6 valence electrons prefer high spin structures, and low spin ground states are derived for the others because valence electron configurations are formed by filling the electron shells 1s/1p/2s/2p based on Pauli's and Hund's rules. Odd-even alternation is observed for both neutral and cationic clusters. Because of the closed electronic shells, the 8- and 10-electron systems are more stable than the others, and the 8-electron species (Li4Ge, Li5Ge+) are more favored than the 10-electron ones (Li6Ge, Li7Ge+). This behavior for Ge is different from C in their doped Li clusters, which can be attributed to the difference in atomic radii. The averaged binding energy plot for neutrals tends to increase slowly with the increasing number of Li atoms, while the same plot for cations shows a maximum at Li5Ge+, which is in good agreement with the mass spectrometry experiment. Atom-in-molecules (AIM) analysis suggests that Li atoms do not bond to one another but through Ge or pseudoatoms, and an essentially ionic character can be attributed to the cluster chemical bonds. An interesting finding is that the larger clusters have the smallest adiabatic ionization energies known so far (IEa approximately 3.5 eV).

Inferring disease and pathway associations of long non-coding RNAs using heterogeneous information network model.

Recent findings from biological experiments demonstrate that long non-coding RNAs (lncRNAs) are actively involved in critical cellular processes and are associated with innumerable diseases. Computational prediction of lncRNA-disease association draws tremendous research attention nowadays. This paper proposes a machine learning model that predicts lncRNA-disease associations using Heterogeneous Information Network (HIN) of lncRNAs and diseases. A Support Vector Machine classifier is developed using the feature set extracted from a meta-path-based parameter, Association Index derived from the HIN. Performance of the model is validated using standard statistical metrics and it generated an AUC value of 0.87, which is better than the existing methods in the literature. Results are further validated using the recent literature and many of the predicted lncRNA-disease associations are identified as actually existing. This paper also proposes an HIN-based methodology to associate lncRNAs with pathways in which they may have biological influence. A case study on the pathway associations of four well-known lncRNAs (HOTAIR, TUG1, NEAT1, and MALAT1) has been conducted. It has been observed that many times the same lncRNA is associated with more than one biologically related pathways. Further exploration is needed to substantiate whether such lncRNAs have any role in determining the pathway interplay. The script and sample data for the model construction is freely available at http://bdbl.nitc.ac.in/LncDisPath/index.html.

Electronic structure of germanium monohydrides Ge(n)H, n = 1-3.

Quantum chemical calculations were applied to investigate the electronic structure of germanium hydrides, Ge(n)H (n = 1, 2, 3), their cations, and anions. Computations using a multiconfigurational quasi-degenerate perturbation approach (MCQDPT2) based on complete active space wave functions (CASSCF), multireference perturbation theory (MRMP2), and density functional theory reveal that Ge(2)H has a (2)B(1) ground state with a doublet-quartet gap of approximately 39 kcal/mol. A quasidegenerate (2)A(1) state has been derived to be 2 kcal/mol above the ground state (MCQDPT2/aug-cc-pVTZ). In the case of the cation Ge(3)H(+) and anion Ge(3)H(-), singlet low-lying electronic states are derived, that is, (1)A' and (1)A(1), respectively. The singlet-triplet energy gap is estimated to 6 kcal/mol for the cation. An "Atoms in Molecules" (AIM) analysis shows a certain positive charge on the Ge(n) (n = 1, 2, 3) unit in its hydrides, in accordance with the NBO analysis. The topologies of the electron density of the germanium hydrides are different from that of the lithium-doped counterparts. On the basis of our electron localization function (ELF) analysis, the Ge-H bond in Ge(2)H is characterized as a three-center-two-electron bond. Some key thermochemical parameters of Ge(n)H have also been derived.

Microfluidic microscopy-assisted label-free approach for cancer screening: automated microfluidic cytology for cancer screening.

Each year, about 7-8 million deaths occur due to cancer around the world. More than half of the cancer-related deaths occur in the less-developed parts of the world. Cancer mortality rate can be reduced with early detection and subsequent treatment of the disease. In this paper, we introduce a microfluidic microscopy-based cost-effective and label-free approach for identification of cancerous cells. We outline a diagnostic framework for the same and detail an instrumentation layout. We have employed classical computer vision techniques such as 2D principal component analysis-based cell type representation followed by support vector machine-based classification. Analogous to criminal face recognition systems implemented with help of surveillance cameras, a signature-based approach for cancerous cell identification using microfluidic microscopy surveillance is demonstrated. Such a platform would facilitate affordable mass screening camps in the developing countries and therefore help decrease cancer mortality rate.

GPURFSCREEN: a GPU based virtual screening tool using random forest classifier.

BACKGROUND: In-silico methods are an integral part of modern drug discovery paradigm. Virtual screening, an in-silico method, is used to refine data models and reduce the chemical space on which wet lab experiments need to be performed. Virtual screening of a ligand data model requires large scale computations, making it a highly time consuming task. This process can be speeded up by implementing parallelized algorithms on a Graphical Processing Unit (GPU). RESULTS: Random Forest is a robust classification algorithm that can be employed in the virtual screening. A ligand based virtual screening tool (GPURFSCREEN) that uses random forests on GPU systems has been proposed and evaluated in this paper. This tool produces optimized results at a lower execution time for large bioassay data sets. The quality of results produced by our tool on GPU is same as that on a regular serial environment. CONCLUSION: Considering the magnitude of data to be screened, the parallelized virtual screening has a significantly lower running time at high throughput. The proposed parallel tool outperforms its serial counterpart by successfully screening billions of molecules in training and prediction phases.