Article
Biochemical Research Methods
Qilong Wu, Sheng-You Huang
Summary: Covalent inhibitors have attracted attention for their long residence time, high binding efficiency, and strong selectivity. The development of computational tools like HCovDock, an efficient docking algorithm for covalent protein-ligand interactions, is valuable for modeling and screening of covalent drugs. HCovDock outperforms seven other state-of-the-art covalent docking programs and exhibits a high success rate in reproducing experimentally observed structures and virtual screening.
BRIEFINGS IN BIOINFORMATICS
(2023)
Article
Biochemical Research Methods
Qilong Wu, Sheng-You Huang
Summary: Covalent inhibitors are highly valued for their long residence time, high binding efficiency, and strong selectivity. The development of computational tools like molecular docking, such as HCovDock, is important for modeling covalent protein-ligand interactions and screening potential drugs. HCovDock shows better performance than other state-of-the-art docking programs and has high success rates in reproducing experimentally observed structures.
BRIEFINGS IN BIOINFORMATICS
(2022)
Article
Multidisciplinary Sciences
Lihua Zhang, Jing Zhang, Qing Nie
Summary: The emergence of single-cell multiomics data provides unprecedented opportunities to study the transcriptional regulatory mechanisms controlling cell identity. DIRECT-NET, a machine-learning method based on gradient boosting, allows us to identify genome-wide cis-regulatory elements (CREs) and their relationship to target genes from single-cell data. DIRECT-NET substantially improves the accuracy of inferring CRE-to-gene relationships and reveals cell subpopulation-specific and dynamic regulatory linkages.
Article
Biochemical Research Methods
Sofia A. Quinodoz, Prashant Bhat, Peter Chovanec, Joanna W. Jachowicz, Noah Ollikainen, Elizabeth Detmar, Elizabeth Soehalim, Mitchell Guttman
Summary: A fundamental question in gene regulation is how cell-type-specific gene expression is influenced by the packaging of DNA within the nucleus. SPRITE is a recently developed method that accurately maps higher-order DNA interactions within the nucleus by cross-linking interacting molecules and using a barcoding method for sequencing. We provide a detailed experimental protocol and an automated computational pipeline for SPRITE.
Article
Chemistry, Multidisciplinary
Stuart T. Johnston, Matthew Faria
Summary: Designing nano-engineered particles for targeted delivery of therapeutic and diagnostic agents is a challenging task. This study presents a machine learning framework that can interpret particle-cell interactions from experimental data, providing insights into the design choices. The framework reveals consistent models of particle-cell interactions for different particle-cell pairs, highlighting the importance of nonlinear saturation effects. Additionally, the framework facilitates quantitative evaluation of particle design choices by providing robust estimates of particle performance.
Article
Biochemistry & Molecular Biology
Pongsakorn Wangkumhang, Matthew Greenfield, Garrett Hellenthal
Summary: We propose fastGLOBETROTTER, a new haplotype-based technique that efficiently identifies, dates, and describes admixture events using genome-wide autosomal data. Simulation results show that fastGLOBETROTTER significantly reduces computation time compared to the related technique GLOBETROTTER, without sacrificing accuracy. Applying fastGLOBETROTTER to a cohort of over 6000 Europeans from 10 countries, we discover previously unreported admixture signals, including multiple periods of admixture related to East Asian or Siberian-like sources in countries north of the Baltic Sea, and admixture related to West Asian, North African, and/or Southern European sources in populations south of the Baltic Sea, dating back to the fall of the Roman Empire. Our approach is scalable and suitable for analyzing large-scale cohorts of genetically homogeneous populations.
Article
Biochemical Research Methods
Charles Homsi, Roshan Elizabeth Rajan, Robin Minati, Edlie St-Hilaire, Eric Bonneil, Simon F. Dufresne, Hugo Wurtele, Alain Verreault, Pierre Thibault
Summary: This article describes a simple filter-aided sample preparation method for extracting and purifying histones, which eliminates tedious steps and is particularly suitable for yeast cells. The method not only improves extraction efficiency, but also inactivates histone-modifying enzymes. The authors demonstrate that the method prevents the common artifact of N-terminal clipping of H3 in yeast cells. It is scalable and enables efficient histone recovery from as few as two million yeast cells, and can be used for analyzing histone modifications in limited fungal clinical isolates.
JOURNAL OF PROTEOME RESEARCH
(2023)
Article
Biotechnology & Applied Microbiology
Xinguo Lu, Fang Liu, Qiumai Miao, Ping Liu, Yan Gao, Keren He
Summary: Our study aimed to reveal functional overlapping patterns in gene modules to elucidate regulatory relationships between overlapping genes and communities, as well as explore cancer formation and progression. We analyzed six cancer datasets and identified three types of gene functional modules for each cancer, with our method outperforming others in terms of identifying distinguishing communities and survival prognostics for patients. In conclusion, overlapping genes play a crucial role in constructing comprehensive carcinogenesis by establishing communication bridges between different specific functional groups.
Review
Biochemistry & Molecular Biology
Calvin K. Voong, James A. Goodrich, Jennifer F. Kugel
Summary: HMGB proteins are small architectural DNA binding proteins that regulate multiple genomic processes, including DNA damage repair and transcription. They can influence disease states, such as cancers and autoimmune diseases, by modulating the local chromatin environment.
Article
Multidisciplinary Sciences
Esha Dutta, Michael A. DeJesus, Nadine Ruecker, Anisha Zaveri, Eun-Ik Koh, Christopher M. Sassetti, Dirk Schnappinger, Thomas R. Ioerger
Summary: Chemical-genetics (C-G) experiments are used to identify interactions between inhibitory compounds and bacterial genes, potentially revealing drug targets or other functionally interacting genes and pathways. By constructing a library of hypomorphic strains, treating them with inhibitory compounds, and using high-throughput sequencing, changes in relative abundance of individual mutants can be quantified. A new statistical method called CGA-LMM is proposed for analyzing C-G data, capturing the dependence of gene abundance in the hypomorph library on increasing drug concentrations through slope coefficients. This method was applied to analyze interactions between Mycobacterium tuberculosis hypomorph libraries and antibiotics, successfully identifying known target genes or expected interactions for the majority of drugs tested.
Article
Biology
Jinhang Wei, Linlin Zhuo, Shiyao Pan, Xinze Lian, Xiaojun Yao, Xiangzheng Fu
Summary: Noncoding RNA (ncRNA), derived from DNA transcription, can participate in gene expression and affect protein synthesis, playing a crucial role in biological processes. Predicting ncRNA-protein interactions is important, and machine learning methods, such as the graph neural network model (GNN), have been developed for this purpose. However, existing models lack a general framework. We propose a GNN-based framework that utilizes topological structure information to predict ncRNA-protein interactions faster and more accurately. We also introduce a new sampling method called HeadTailTransfer to mitigate the limitations of smaller and larger datasets. Experimental results demonstrate the effectiveness of our approach.
COMPUTERS IN BIOLOGY AND MEDICINE
(2023)
Article
Multidisciplinary Sciences
Esha Dutta, Michael A. DeJesus, Nadine Ruecker, Anisha Zaveri, Eun-Ik Koh, Christopher M. Sassetti, Dirk Schnappinger, Thomas R. Ioerger
Summary: Chemical-genetics experiments can identify interactions between inhibitory compounds and bacterial genes, revealing drug targets or functionally interacting genes. By using Linear Mixed Models, a statistical method known as CGA-LMM can analyze C-G data and detect candidate gene interactions based on their abundance changes with increasing drug concentrations.
Article
Computer Science, Information Systems
Xuemin Zhao, Ran Duan, Shaowen Yao
Summary: Topologically associated domains (TADs) are essential units in chromatin's three-dimensional organization, influencing various biological processes. This study proposes an optimization method for TAD identification using empirical mode decomposition and compares its results with five commonly used TAD detection methods. The proposed method demonstrates universality and efficiency, highlighting its potential as a valuable tool in TAD identification.
Article
Biochemistry & Molecular Biology
Haiyan Gong, Yi Yang, Xiaotong Zhang, Minghong Li, Sichen Zhang, Yang Chen
Summary: CASPIAN is a method based on spatial density clustering algorithm for accurately identifying chromatin TAD boundaries and enriching factors related to gene expression. It is implemented using HDBSCAN and adapts well to Hi-C contact matrices of different resolutions.
COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL
(2022)
Article
Chemistry, Multidisciplinary
Roderick Corstiaan Abraham Versloot, Sabine Angenieta Paulina Straathof, Gemma Stouwie, Matthijs Jonathan Tadema, Giovanni Maglia
Summary: Biological nanopores are sensitive single-molecule sensors for proteins and peptides. This research shows that modifying nanopores with an acidic-aromatic sensing region can dramatically increase the discrimination of peptides in the nanopore at acidic pH. The capture mechanisms and resolutions of two beta-barrel nanopores, aerolysin and cytotoxin K, were found to differ.
Letter
Dermatology
Donna M. Brennan-Crispi, Maxwell Frankfurter, Christina Murphy, Emily Sheng, Mingang Xu, Edward E. Morrisey, Sarah E. Millar, Thomas H. Leung
JOURNAL OF INVESTIGATIVE DERMATOLOGY
(2022)