Single-Cell RNA Sequencing Reveals an Immune Landscape of CD4+ T Cells in Coronary Culprit Plaques With Acute Coronary Syndrome in Humans-Brief Report.

BACKGROUND: Acute coronary syndrome (ACS) involves plaque-related thrombosis, causing primary ischemic cardiomyopathy or lethal arrhythmia. We previously demonstrated a unique immune landscape of myeloid cells in the culprit plaques causing ACS by using single-cell RNA sequencing. Here, we aimed to characterize T cells in a single-cell level, assess clonal expansion of T cells, and find a therapeutic target to prevent ACS. METHODS: We obtained the culprit lesion plaques from 4 patients with chronic coronary syndrome (chronic coronary syndrome plaques) and the culprit lesion plaques from 3 patients with ACS (ACS plaques) who were candidates for percutaneous coronary intervention with directional coronary atherectomy. Live CD45+ immune cells were sorted from each pooled plaque samples and applied to the 10× platform for single-cell RNA sequencing analysis. We also extracted RNA from other 3 ACS plaque samples and conducted unbiased TCR (T-cell receptor) repertoire analysis. RESULTS: CD4+ T cells were divided into 5 distinct clusters: effector, naive, cytotoxic, CCR7+ (C-C chemokine receptor type 7) central memory, and FOXP3 (forkhead box P3)+ regulatory CD4+ T cells. The proportion of central memory CD4+ T cells was higher in the ACS plaques. Correspondingly, dendritic cells also tended to express more HLAs (human leukocyte antigens) and costimulatory molecules in the ACS plaques. The velocity analysis suggested the differentiation flow from central memory CD4+ T cells into effector CD4+ T cells and that from naive CD4+ T cells into central memory CD4+ T cells in the ACS plaques, which were not observed in the chronic coronary syndrome plaques. The bulk repertoire analysis revealed clonal expansion of TCRs in each patient with ACS and suggested that several peptides in the ACS plaques work as antigens and induced clonal expansion of CD4+ T cells. CONCLUSIONS: For the first time, we revealed single cell-level characteristics of CD4+ T cells in patients with ACS. CD4+ T cells could be therapeutic targets of ACS. REGISTRATION: URL: https://upload.umin.ac.jp/cgi-open-bin/icdr_e/ctr_view.cgi?recptno=R000046521; Unique identifier: UMIN000040747.

Estimation of important binding sites in compounds that interact with proteins.

Proteins are one of the important substances in understanding biological activity, and many of them express the function by binding to other proteins or small molecules (ligands) on the molecular surface. This interaction often occurs in the hollows (pockets) on the molecular surface of the protein. It is known that when pockets are similar in structure and physical properties, they are likely to express similar functions and to bind similar ligands. Therefore, exploring the similarity of the structure and physical properties in pockets is very useful because it leads to the discovery of new ligands that are likely to bind. In addition, exploring the important structure when binding to the protein significant spot in the ligand will provide useful knowledge for the development of new ligands. In this study, we propose a method to search for proteins containing pockets that are structurally and physically similar to significant spot in the pocket of the analyzed protein, and to extract significant spots in the ligands that bind to them. We use feature points as data. Feature points are the 3-dimensional points that are extracted from 3D structure data of proteins with feature values quantifying hydrophobicity and electrostatic potential. The corresponding feature points are extracted by comparing structurally and physically the pockets of the search target proteins with the significant spot of the analyzed protein. By evaluating the similarity based on the comparison results of the feature values given to the extracted feature points, we search for proteins that are similar to the analyzed protein. From the ligands that bind to the searched proteins, atoms that are near the protein pocket and similar to the atoms in ligand binding to the analyzed protein are extracted. The site constituted by the extracted atoms is defined as a significant spot in the ligand. As a result of classifying ligands binding to the protein by using the extracted significant spot in the ligand, the effectiveness of the proposed method was confirmed.

Protein-protein interaction extraction with feature selection by evaluating contribution levels of groups consisting of related features.

BACKGROUND: Protein-protein interaction (PPI) extraction from published scientific articles is one key issue in biological research due to its importance in grasping biological processes. Despite considerable advances of recent research in automatic PPI extraction from articles, demand remains to enhance the performance of the existing methods. RESULTS: Our feature-based method incorporates the strength of many kinds of diverse features, such as lexical and word context features derived from sentences, syntactic features derived from parse trees, and features using existing patterns to extract PPIs automatically from articles. Among these abundant features, we assemble the related features into four groups and define the contribution level (CL) for each group, which consists of related features. Our method consists of two steps. First, we divide the training set into subsets based on the structure of the sentence and the existence of significant keywords (SKs) and apply the sentence patterns given in advance to each subset. Second, we automatically perform feature selection based on the CL values of the four groups that consist of related features and the k-nearest neighbor algorithm (k-NN) through three approaches: (1) focusing on the group with the best contribution level (BEST1G); (2) unoptimized combination of three groups with the best contribution levels (U3G); (3) optimized combination of two groups with the best contribution levels (O2G). CONCLUSIONS: Our method outperforms other state-of-the-art PPI extraction systems in terms of F-score on the HPRD50 corpus and achieves promising results that are comparable with these PPI extraction systems on other corpora. Further, our method always obtains the best F-score on all the corpora than when using k-NN only without exploiting the CLs of the groups of related features.

A method of searching for related literature on protein structure analysis by considering a user's intention.

BACKGROUND: In recent years, with advances in techniques for protein structure analysis, the knowledge about protein structure and function has been published in a vast number of articles. A method to search for specific publications from such a large pool of articles is needed. In this paper, we propose a method to search for related articles on protein structure analysis by using an article itself as a query. RESULTS: Each article is represented as a set of concepts in the proposed method. Then, by using similarities among concepts formulated from databases such as Gene Ontology, similarities between articles are evaluated. In this framework, the desired search results vary depending on the user's search intention because a variety of information is included in a single article. Therefore, the proposed method provides not only one input article (primary article) but also additional articles related to it as an input query to determine the search intention of the user, based on the relationship between two query articles. In other words, based on the concepts contained in the input article and additional articles, we actualize a relevant literature search that considers user intention by varying the degree of attention given to each concept and modifying the concept hierarchy graph. CONCLUSIONS: We performed an experiment to retrieve relevant papers from articles on protein structure analysis registered in the Protein Data Bank by using three query datasets. The experimental results yielded search results with better accuracy than when user intention was not considered, confirming the effectiveness of the proposed method.

Extraction of Protein-Protein Interaction from Scientific Articles by Predicting Dominant Keywords.

For the automatic extraction of protein-protein interaction information from scientific articles, a machine learning approach is useful. The classifier is generated from training data represented using several features to decide whether a protein pair in each sentence has an interaction. Such a specific keyword that is directly related to interaction as "bind" or "interact" plays an important role for training classifiers. We call it a dominant keyword that affects the capability of the classifier. Although it is important to identify the dominant keywords, whether a keyword is dominant depends on the context in which it occurs. Therefore, we propose a method for predicting whether a keyword is dominant for each instance. In this method, a keyword that derives imbalanced classification results is tentatively assumed to be a dominant keyword initially. Then the classifiers are separately trained from the instance with and without the assumed dominant keywords. The validity of the assumed dominant keyword is evaluated based on the classification results of the generated classifiers. The assumption is updated by the evaluation result. Repeating this process increases the prediction accuracy of the dominant keyword. Our experimental results using five corpora show the effectiveness of our proposed method with dominant keyword prediction.

Interaction site prediction by structural similarity to neighboring clusters in protein-protein interaction networks.

BACKGROUND: Recently, revealing the function of proteins with protein-protein interaction (PPI) networks is regarded as one of important issues in bioinformatics. With the development of experimental methods such as the yeast two-hybrid method, the data of protein interaction have been increasing extremely. Many databases dealing with these data comprehensively have been constructed and applied to analyzing PPI networks. However, few research on prediction interaction sites using both PPI networks and the 3D protein structures complementarily has explored. RESULTS: We propose a method of predicting interaction sites in proteins with unknown function by using both of PPI networks and protein structures. For a protein with unknown function as a target, several clusters are extracted from the neighboring proteins based on their structural similarity. Then, interaction sites are predicted by extracting similar sites from the group of a protein cluster and the target protein. Moreover, the proposed method can improve the prediction accuracy by introducing repetitive prediction process. CONCLUSIONS: The proposed method has been applied to small scale dataset, then the effectiveness of the method has been confirmed. The challenge will now be to apply the method to large-scale datasets.

A method for supporting retrieval of articles on protein structure analysis considering users' intention.

BACKGROUND: In recent years, information about protein structure and function is described in a large amount of articles. However, a naive full-text search by specific keywords often fails to find desired articles, because the articles involve the ambiguous and complicated concepts that cannot be described with uniform representation. For retrieving articles on protein structure and function, it is important to consider the relevance between structural and/or functional concepts by identifying the user's intention. RESULTS: We introduce a scheme of evaluating relevance between articles based on various biological databases and ontologies on structures and functions of proteins. The relevance, which is defined as a path length between concepts on hierarchies, is modified adaptively based on additional articles as a query in order to reflect the user's intention. Also we implemented the retrieval system, in which the user can input some articles as a query and the related articles are retrieved and displayed on the 2D map. CONCLUSIONS: The effectiveness of the proposed system was confirmed experimentally by having shown that the users can obtain easily highly related articles which reflect their intention.