BigBioMdata

проект КП-06-Н37/24 Иновативна интегрирана платформа за интелигентно управление и анализ на потоци големи данни за биомедицински научни изследвания

Публикации

Списък с публикации по проекта

  1. Gancheva V., Platform for big biomedical data streams management and analytics, International Journal of Circuits, Systems and Signal Processing, 2020, 14, pp. 580–588, Scopus, SJR=0.16, Q4. https://www.naun.org/main/NAUN/circuitssystemssignal/2020/b502005-bnv.pdf
  2. Gancheva V., Jongov T., Medical X-ray Image Classification Method Based on Convolutional Neural Networks for Diagnosing COVID-19, accepted for publishing, International Journal Bioautomation, SJR=0.178, Q4.
  3. Draganov I., Gancheva V. Unsharp Masking with Local Adaptive Contrast Enhancement of Medical Images. In: Su R., Zhang YD., Liu H. (eds), Lecture Notes in Electrical Engineering, LNEE, vol. 784, 2022, pp. 354–363. Springer, https://doi.org/10.1007/978-981-16-3880-0_37, Scopus, SJR=0.13, Q4. https://link.springer.com/chapter/10.1007/978-981-16-3880-0_37
  4. Vetova S., Big heterogeneous data integration and analysis, AIP Conference Proceedings, 2021, 2333, 030007, https://doi.org/10.1063/5.0043627, Scopus, SJR=0.18. https://aip.scitation.org/doi/10.1063/5.0043627
  5. Vetova S., Big data workflow platforms for science, AIP Conference Proceedings, 2021, 2333, 030008, https://doi.org/10.1063/5.0043625, Scopus, SJR=0.18. https://aip.scitation.org/doi/10.1063/5.0043625
  6. Gancheva V., Knowledge Discovery Based on Data Analytics and Visualization Supporting Precision Medicine, Proceedings of 2nd International Conference on Mathematics and Computers in Science and Engineering, MACISE 2020, pp. 102–105, 9195603, DOI: 10.1109/MACISE49704.2020.00024, Scopus. https://ieeexplore.ieee.org/document/9195603
  7. Ko S.-H., Gancheva, V., An Approach for Parallel Reading in Multiple Sequence Alignment, Proceedings of International Conference Automatics and Informatics, ICAI 2020, DOI: 10.1109/ICAI50593.2020.9311347, Scopus. https://ieeexplore.ieee.org/document/9311347
  8. Vetova S., Biomedical Data Integration and Innovations Concept, Applications of Mathematics in Engineering and Economics (AMEE’21), 2021, AIP Conference Proceedings, accepted for publishing.
  9. Vetova S., Biomedical Image Classification Algorithms Evaluation, Applications of Mathematics in Engineering and Economics (AMEE’21), 2021, AIP Conference Proceeding, accepted for publishing.
  10. Vetova S., Determination of Accuracy and Probability in the Analysis of Large-Scale Biomedical Data, Applications of Mathematics in Engineering and Economics (AMEE’21), 2021, AIP Conference Proceedings, accepted for publishing.
  11. Vetova S., Wavelet Analysis for Biomedical Data, Applications of Mathematics in Engineering and Economics (AMEE’21), 2021, AIP Conference Proceedings, accepted for publishing.
  12. Gancheva V., Parallel Multithreaded Medical Images Filtering, IEEE International Conference on Computational Science and Computational Intelligence (CSCI’21): December 15-17, 2021; Las Vegas, USA), accepted for publishing, Scopus.
  13. Draganov I., Gancheva V. Optimal Bilateral Filtering of CT Images, IEEE International Conference on Computational Science and Computational Intelligence (CSCI’21): December 15-17, 2021; Las Vegas, USA), accepted for publishing, Scopus.

Резюмета на представени резултати на научни форуми

1. Ko S.-H., Gancheva, V., An Approach for Parallel Reading in Multiple Sequence Alignment, International Conference Automatics and Informatics, ICAI, 1-3 October 2020, Varna, Bulgaria.

Abstract: We propose an approach for faster file reading of multiple sequence alignment input through the use of MPI-I/O over a subset of MPI cores. The idea is to let a subset of MPI cores to perform the I/O operation and locally broadcast to individual neighbors so that the code is less sensitive to the stability of the parallel file system. It is achieved by creating a number of subgroups under a global MPI communicator. The size of each subgroup and the buffer size of each reading operation are tuned through the synthetic benchmark. We verify the performance of our approach by comparing it with the traditional way of “sequential file reading and global broadcast”, and apply it to the MPI version of multiple sequence alignment software ClustalW. In the production run over 8192 BlueGene/Q cores, the current approach provides 6.8 times speed-up than the original ClustalW-MPI implementation.

2. Draganov I., Gancheva V. Unsharp Masking with Local Adaptive Contrast Enhancement of Medical Images, International Conference on Medical Imaging and Computer-Aided Diagnosis, MICAD, 25-26 Mart 2021, Birmingham, UK.

Abstract: In this paper we present a generalized algorithm for unsharp masking of medical images which takes as one of its inputs a high contrast image underwent local adaptive contrast enhancement. Selection of optimal values of the number of histogram bins, processing window size and intensity lower and upper limits in iterative manner is part of applying Contrast Limited Adaptive Histogram Equalization (CLAHE). Experimental results reveal higher quality of the output images both in terms of root mean square contrast and sharpness. Achieved quality, both visually and quantitatively, is compared to that from the Adaptive Histogram Equalization (AHE) algorithm, limited histogram stretching and ordinary histogram equalization which proves its applicability. The algorithm is considered appropriate for processing a number of types of images, such as CT, X-ray, etc.

3. Vetova S., Biomedical Data Integration and Innovations Concept, 47th International Conference Applications of Mathematics in Engineering and Economics, 7 – 13 June 2021, Sozopol, Bulgaria.

Abstract: The presented paper deals with data integration, filtering, sorting and aggregation of Covid-19 data. The data file contains eight data fields and for the design of the proposed model for data integration and processing each of them is configured in data type, format and field length. Being designed in the Talend Open Studio, the model concerns the performance of six main tasks: data integration, data filtering, data sorting, data aggregation, data sorting after the process of aggregation and result display.

4. Vetova S., Biomedical Image Classification Algorithms Evaluation, 47th International Conference Applications of Mathematics in Engineering and Economics, 7 – 13 June 2021, Sozopol, Bulgaria.

Abstract: The object of the following paper is a concept for biomedical image classification. On its base, two alorithms for Covid-19 image classification are designed. For the image decomposition and feature vectors a wavelet transform is used and in particular the Dual Tree Complex Wavelet Transform. For the first algorithm, wavelet transform is performed at the third level and the second algorithm is designed at the second transform level. A comparative analysis on both algorithms is accomplished based on the evaluation metrics accuracy, precision and recall. The results are depicted graphically.

5. Vetova S., Determination of Accuracy and Probability in the Analysis of Large-Scale Biomedical Data, 47th International Conference Applications of Mathematics in Engineering and Economics, 7 – 13 June 2021, Sozopol, Bulgaria.

Abstract: The paper below introduces an image classification workflow concept for biomedical purposes and present the process of determination of accuracy and probabilities in the analysis of large-scale biomedical data. The workflow is based on image features extraction through deep network embedding. The distance computation in the workflow is accomplished using Cosine distance. The experiments are conducted on the base of the Logistic Regression, Random Forest and Naïve Bayes and are directed on accuracy and probability in the analysis of large-scale data. The performed analysis shows the advantage of the Logistic Regression results which are considered to be the most reliable compared to the results generated through the Random Forest and Naïve Bayes.

6. Vetova S., Wavelet Analysis for Biomedical Data, 47th International Conference Applications of Mathematics in Engineering and Economics, 7 – 13 June 2021, Sozopol, Bulgaria.

Abstract: The following aper deals with a concept for biomedical image classification based on the wavelet analysis. For the goal of the concept implementation Covid-19 image database is applied. For image analysis and description, the Dual- Tree Complex Wavelet Transform is applied in combination with the Euclidean distance to compute the similarity distance. The proposed concept is evaluated through accuracy, precision and recall metric. The results are presented and analysed showing the advantages of the concept for biomedical image classification and its application.

7. Anastas Pashov, Shina Pashova, Peter Petrov, Formal Representation of the Repertoire of IgM Antibody Specificities, The Workshop “Mathematics of Life” (MoL2021), 13-16 September 2021, Hisarya, Bulgaria.

Abstract: Testing of individual antibody reactivtities as an interrogation of the immune history of an individual is routine practice. Recently it has become increasingly evident that the global view of the repertoire of antibodies may provide information about the internal environment beyond the sum of the individual immune responses and the immune history. As part of the multi- omics endeavor to capture a holistic image of the biological systems, repertoire studies develop in two directions – repertoire sequencing (RepSeq) and functional probing of the repertoire with arrays of diverse structures (e. g. peptides or glycans, referred to as igome). We used the igome approach based on high throughput probing of the human IgM repertoire by selection of mimotopes from phage display random peptide library followed by next generation sequencing. This technique yields 105 106 different sequences each a target for at least one antibody in the repertoire. These sequences fit the same binding pocket as a putative nominal epitope for a given anti- body without representing exactly the same sequence or structure, hence – mimotopes. To analyze bioinformatically the igome image of the repertoire first a suitable metric reflecting the mimotope properties was selected using ROC curves.  Longest common subsequence proved most suitable although it predicted only about 80% of the mimotopes of a given monoclonal anti- body.  Next the igome was represented as a graph of sequences connected if their longest common subsequence exceeds a conservative threshold of 4 out of 7 residues. In our first analysis, this graph helped us define neighbor- hoods of sequences which contain predominantly mimotopes of the repertoire in healthy donors vs neighborhoods representing mostly mimotopes of the repertoire of patients with anti-phospholipid syndrome. A key feature of the IgM mimotopes proved to be them mirroring sequences from the binding site (paratope) of other antibodies. This is a confirmation of the controversial theory of the antibody networks (idiotypy). A correlation between idiotopes (epitopes within the paratopes of other antibodies) and biological function was found. These initial studies attest to the igome’s potential to provide a new class of biomarkers.

8. Anastas Pashov, Peter Petrov, Topological Approach for a Global Description of the Antibody Repertoire, Formal Representation of the Repertoire of IgM Antibody Specificities, The Workshop “Mathematics of Life” (MoL2021), 13-16 September 2021, Hisarya, Bulgaria.

Abstract: We are interested in characterizing the antibody-antigen interactions in a high throughput “omics” context. Epitopes are small domains (10 to 22 residues) of the protein molecules identified as targets of binding by specific antibodies. Our approach is based on the analysis of large libraries of potential epitopes or their simulacra – short peptide mimotopes. We propose to apply persistent homology after formalizing the relations between the epitopes (mimotopes) using the longest common subsequence metric. The concept of the shape of the data cloud, and how to analyze it rigorously using tools from topology will be discussed. Our aim is to obtain some qualitative description of the data cloud of epitopes with more than 3×106 objects, applicable to studies of antibody-antigen recognition, antibody (poly) specificity, and immune networks.