• Korean Journal of Cognitive Science
• /
• v.28 no.4
• /
• pp.299-314
• /
• 2017

Predicting compound-protein interactions in-silico is significant for the drug discovery. In this paper, we propose an scalable machine learning model to predict compound-protein interaction. The key idea of this scalable machine learning model is the architecture of pairwise neural network model and feature embedding method from the raw data, especially for protein. This method automatically extracts the features without additional knowledge of compound and protein. Also, the pairwise architecture elevate the expressiveness and compact dimension of feature by preventing biased learning from occurring due to the dimension and type of features. Through the 5-fold cross validation results on large scale database show that pairwise neural network improves the performance of predicting compound-protein interaction compared to previous prediction models.

• Journal of the Korean Magnetic Resonance Society
• /
• v.20 no.1
• /
• pp.22-26
• /
• 2016

FANCJ is a DNA helicase which contributes genome stability by resolving G-quadruplex DNA from 5' to 3' direction. In addition to main ATPase helicase core, FANCJ has the protein binding region at its C-terminal part. BRCA1 and BLM are the binding partner of FANCJ and these protein-protein interactions contribute genomic stability and the proper response to replication stress. As the first attempt for studying FANCJ-BLM interaction, we prepared BLM binding region of FANCJ and characterized with CD and NMR spectroscopy. FANCJ (881-941) with N-ter 6xHis was purified as the oligomer. Secondary structure prediction based on CD data revealed that FANCJ (881-941) composed with ${\beta}$ sheet, turn and coils.$^1H-^{15}N$ HSQC spectra showed nonhomogeneous peak intensities with less number of peaks comparing than the number of amino acids in the construct. It indicated that optimization should be necessary for detailed further structural studies.

• Korean Journal of Veterinary Research
• /
• v.24 no.1
• /
• pp.24-30
• /
• 1984

The protective effects of high levels of dietary iron, copper, zinc or protein on lead toxicity were studied In chicks. Growth retardation, reduction of feed intake, anemia and accumulation of lead in the bone and kidney were observed in chicks fed a diet containing 500mg lead as chloride per kg of feed for 42 days. Early changes due to ingested lead were inhibition of red blood cell ${\delta}$-aminolevulinic acid dehydrase at all doses and no effect of iron, copper, zinc or protein addition were observed. Tibia lead accumulation was reduced in chicks receiving additional dietary iron or zinc compared to the lead only group but increased in chicks given supplementary protein. Decreased body weight gain was overcome by supplementary zinc or protein in chicks fed lead but not by supplementary iron. Overall the results of this study show that lead poisoning can be partly reduced by providing supplementary iron, zinc or protein, but the interaction of these element remained to be elucidated.

• Journal of Integrative Natural Science
• /
• v.10 no.3
• /
• pp.125-130
• /
• 2017

Glucagon like pepide-2, one of the GLPs, is involved in various metabolic functions in the gastrointestinal tract. It plays a major role in the regulation of mucosal epithelium and the intestinal crypt cell proliferation. Because of their therapeutic importance towards the diseases in the gastrointestinal tract, it becomes necessary to study their interaction with its receptor, GLP-2R. In this study, we have developed protein-protein docking complexes of GLP-2 - GLP-2 receptor. Homology models of GLP-2 are developed, and a reliable model out of the predicted models was selected after model validation. The model was bound with the receptor, to study the important interactions of the complex. This study could be useful in developing novel and potent drugs for the diseases related with GLP-2.

• Proceedings of the Korean Society of Applied Pharmacology
• /
• 2008.04a
• /
• pp.77-86
• /
• 2008

It's been known that overexpression of the oncoprotein Her2 (eu/ErbB2), transmembrane receptor protein, occurs in human breast cancer. Her2-positive breast cancer patients who have Her2 overexpression show less therapeutic efficacy with enhanced metathesis and increased resistance to chemotherapy. So far, a humanized monoclonal antibody against Her2 protein called Herceptin is the only drug approved by Food and Drug Administration for treatment of Her2-overexpressing breast tumors. However, antibody therapy of Herceptin may not be ideal method for therapeutic intervention of Her2 protein expression. The therapeutic intervention of Her2 protein expression may be more efficiently achieved by inhibiting the expression of Her2 gene rather than by down-regulating the Her2 protein already overexpressed. Here, we found that the interaction of two proteins of ESX (an epithelial-restricted transcription factor) and DRIP130/CRSP130/Sur2 (a Ras-linked subunit of human mediator complexes) mediates the expression of Her2 gene. The association of ESX with Sur2 is mediated by a small hydrophobic face of 8-amino acid helix in ESX, suggesting that the ESX-Sur2 interaction can be a new novel target for Her2-positive cancer. The process to develop potent ESX-Sur2 interaction inhibitors targeting for Her2-positive cancer therapeutics will be discussed.

• BMB Reports
• /
• v.35 no.6
• /
• pp.562-567
• /
• 2002

In order to find the cellular interaction factors of the Heliothis armigera nuclear polyhedrosis virus capsid protein VP39, a Heliothis armigera cell cDNA library was constructed. Then VP39 was used as bait. The host actin gene was isolated from the cDNA library with the yeast two-hybrid system. This demonstrated that VP39 could interact with its host actin in yeast. In order to corroborate this interaction in vivo, the vp39 gene was fused with the green fluorescent protein gene in plasmid pEGFP39. The fusion protein was expressed in the Hz-AM1 cells under the control of the Autographa californica multiple nucleopolyhedrovirus immediate early gene promoter. The host actin was labeled specifically by the red fluorescence substance, tetramethy rhodamine isothicyanete-phalloidin. Observation under a fluorescence microscopy showed that VP39, which was indicated by green fluorescence, began to appear in the cells 6 h after being transfected with pEGFP39. Red actin cables were also formed in the cytoplasm at the same time. Actin was aggregated in the nucleus 9 h after the transfection. The green and red fluorescence always appeared in the same location of the cells, which demonstrated that VP39 could combine with the host actin. Such a combination would result in the actin skeleton rearrangement.

• Journal of Microbiology and Biotechnology
• /
• v.31 no.8
• /
• pp.1183-1189
• /
• 2021

Autodisplay of a multimeric protein complex on a cell surface is limited by intrinsic factors such as the types and orientations of anchor modules. Moreover, improper folding of proteins to be displayed often hinders functional cell surface display. While overcoming these drawbacks, we ultimately extended the applicability of the autodisplay platform to the display of a protein complex. We designed and constructed a cell surface attachment (CSA) system that uses a non-covalent protein-protein interaction. We employed the high-affinity interaction mediated by an orthogonal cohesin-dockerin (Coh-Doc) pair from Archaeoglobus fulgidus to build the CSA system. Then, we validated the orthogonal Coh-Doc binding by attaching a monomeric red fluorescent protein to the cell surface. In addition, we evaluated the functional anchoring of proteins fused with the Doc module to the autodisplayed Coh module on the surface of Escherichia coli. The designed CSA system was applied to create a functional attachment of dimeric α-neoagarobiose hydrolase to the surface of E. coli cells.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.3
• /
• pp.771-791
• /
• 2022

Protein-protein interaction (PPI) extraction from original text is important for revealing the molecular mechanism of biological processes. With the rapid growth of biomedical literature, manually extracting PPI has become more time-consuming and laborious. Therefore, the automatic PPI extraction from the raw literature through natural language processing technology has attracted the attention of the majority of researchers. We propose a PPI extraction model based on the large pre-trained language model and adversarial training. It enhances the learning of semantic and syntactic features using BioBERT pre-trained weights, which are built on large-scale domain corpora, and adversarial perturbations are applied to the embedding layer to improve the robustness of the model. Experimental results showed that the proposed model achieved the highest F1 scores (83.93% and 90.31%) on two corpora with large sample sizes, namely, AIMed and BioInfer, respectively, compared with the previous method. It also achieved comparable performance on three corpora with small sample sizes, namely, HPRD50, IEPA, and LLL.

• Genomics & Informatics
• /
• v.21 no.1
• /
• pp.6.1-6.8
• /
• 2023

Glaucoma is the second leading cause of irreversible blindness, and primary open-angle glaucoma (POAG) is the most common type. Due to inadequate diagnosis, treatment is often not administered until symptoms occur. Hence, approaches enabling earlier prediction or diagnosis of POAG are necessary. We aimed to identify novel drugs for glaucoma through bioinformatics and network analysis. Data from 36 samples, obtained from the trabecular meshwork of healthy individuals and patients with POAG, were acquired from a dataset. Next, differentially expressed genes (DEGs) were identified to construct a protein-protein interaction (PPI) network. In both stages, the genes were enriched by studying the critical biological processes and pathways related to POAG. Finally, a drug-gene network was constructed, and novel drugs for POAG treatment were proposed. Genes with p < 0.01 and |log fold change| > 0.3 (1,350 genes) were considered DEGs and utilized to construct a PPI network. Enrichment analysis yielded several key pathways that were upregulated or downregulated. For example, extracellular matrix organization, the immune system, neutrophil degranulation, and cytokine signaling were upregulated among immune pathways, while signal transduction, the immune system, extracellular matrix organization, and receptor tyrosine kinase signaling were downregulated. Finally, novel drugs including metformin hydrochloride, ixazomib citrate, and cisplatin warrant further analysis of their potential roles in POAG treatment. The candidate drugs identified in this computational analysis require in vitro and in vivo validation to confirm their effectiveness in POAG treatment. This may pave the way for understanding life-threatening disorders such as cancer.

• Proceedings of the Korea Contents Association Conference
• /
• 2004.11a
• /
• pp.511-515
• /
• 2004

In the post-genomic era, researches on proteins as well as genes have been increasingly required. Particularly, work on protein-protein interaction and protein network construction have been recently establishing. Most biologists publish their research results through papers or other media. However, biologists do not use the information effectively, since the published research results are very large. As the growth of internet, it becomes easy to access very large research results. It is significantly important to extract information with a biological meaning from varisous media. Therefore, in this research, we efficiently extract protein-protein interaction information from many open papers or other media and construct the database of the extracted information. We build a protein network from the established database and then design and implement various pathway analysis algorithms which find biological meaning from the protein network.