An epidemic of monkeypox occurred in May 2023 that led to a disease named mpox displaying symptoms such as fever, enlarged lymph nodes, and rash. This ailment is usually mild and self-limited, but severe cases can be harrowing and may cause lifetime scarring. In response to the outbreak, researchers from Mount Sinai and the Carlos III Health Institute (ICI) in Madrid, Spain, worked together to investigate the genetic makeup of the monkeypox virus (MPXV). They mainly studied subclade IIb strains of the virus, helping them to understand the convolutions in viral genes that influence virus behavior, thereby offering strategies for intervention.

The concept of “genomic accordion” refers to periodic expansions and contractions within the genome of monkeypox virus (MPXV), especially in low-complexity regions (LCRs). Adaptive evolution for this virus depends on these variants, which are drivers of gene expression changes or modifications, including short tandem repeats.  

Contrarily, researchers believe that differences in LCRs rather than single-nucleotide polymorphisms (SNPs) may explain the distinct epidemiology of subclade IIb MPXV strains. MPXV’s genetic pathways that underlie its unique transmission dynamics and pathogenicity have been illuminated through precise LCR resolution and repeat length analysis.

Continue Reading

Single-cell transcriptomics has been generating large-scale information, improving our knowledge of cellular processes across different tissues, and facilitating medication discovery, diagnosis, and prognosis. But sorting through all of this data has proven to be an enormous undertaking that frequently takes weeks or months. 

The vast amount of data generated—between hundreds of gigabytes and tens of terabytes—that necessitates a significant amount of time for analysis is the cause of this bottleneck. Biologists also have a challenging learning curve due to the complex set of stages involved in data processing, which needs a variety of software programs. Furthermore, because data analysis in this field is iterative, it takes a profound understanding of biology to develop pertinent questions, carry out analyses, evaluate findings, and more ideas. 

In order to tackle these issues, scientists from the University of California introduced Bioinformatics Copilot 1.0, a software that utilizes a broad language model. With the help of an easy-to-use natural language interface, users may analyze data without needing to be proficient in programming languages like Python or R. It is designed to work on multiple operating systems, including Windows, Linux, and Mac. Significantly, it makes local data analysis easier and guarantees compliance with strict data management guidelines governing the use of patient samples in healthcare and research facilities.

Continue Reading

Just think about yourself as a physician examining an x-ray. You see an area that seems out of the ordinary, but you are unsure of exactly where it starts and ends. At times, there may be more to the story than what current AI models can provide by way of a single response. This is where Tyche comes in, a groundbreaking framework developed by a team at MIT and the Broad Institute to unravel the inherent indeterminacy in medical image segmentation.

Tyche differs from conventional AI systems that provide only one answer by generating a range of possible answers instead. It’s as if many expert consultants were whispering different versions into your ear to enable you to arrive at better decisions.

The Achilles’ Heel of Traditional Segmentation

Medical imaging analysis, ranging from X-rays to MRIs, usually requires segmentation, which is the identification and delineation of specific structures within an image, such as tumors or organs. This process traditionally has been done manually; it’s slow and error-prone. There was some hope in AI, though, promising automation and increased accuracy.

Continue Reading

Imagine a teeming city where proteins are the movers and shakers. Each protein intertwines with others in partnerships that are vital to every cellular process. Nevertheless, unraveling these interactions, which are called protein-protein interactions (PPIs), forms the basis for both biology and drug development. Detecting PPIs has been akin to overhearing hushed conversations in a crowded room.

A recent study by Harvard Medical School researchers makes significant strides toward understanding the landscape of protein interaction. In this research, SPOC, a powerful classifier, was introduced, which improves the highly accurate identification of true PPIs compared to predictions made by AlphaFold-Multimer (AF-M), an innovative protein structure prediction tool. Predictions generated by the SPOC can be accessed and evaluated on the Predictomes database (predictomes.org). It serves as a repository for protein-protein interaction predictions, allowing users to explore and score their own predictions using SPOC.

Continue Reading

The effective paradigm of molecular linkers in drug discovery is crucial for obtaining relevant candidate molecules in early-stage drug development. In this study, researchers present DiffLinker, a three-dimensional conditional diffusion model for molecular linker design that is E(3)-equivariant. A synthetically available model called DiffLinker generates molecules from disconnected fragments that include all of the original fragments. It can automatically count the number of atoms in the linker and its attachment locations to the input fragments. It can link an infinite number of fragments. 

On benchmark datasets, the model performs better than other approaches, producing more varied and readable molecules. It can effectively generate legitimate linkers conditioned on target protein pockets, according to experimental testing conducted in practical applications.

Continue Reading

Understanding molecular mechanisms requires an understanding of protein interactions with nucleic acids. Researchers can query preprocessed and clustered structural data, analyze the data, and draw conclusions about protein interactions based on homology using the PPI3D web server. All interactions for proteins homologous to the query, interactions between two proteins or their homologs, and interactions within a particular PDB entry are the three interaction exploration options that the server provides. 

Protein annotations, structures, interface residues, and contact surface regions are all included in the interactive study. All contact interfaces and binding sites from PDB are included in the weekly updated PPI3D database, which is grouped according to structural similarity and protein sequence. For proteins that are extensively investigated, this prevents repetitive information by producing non-redundant datasets without sacrificing different interaction types.

Continue Reading

Researchers from the Institute of Bioengineering, Lausanne, Switzerland, introduce GEMLI, a powerful computational tool that enables robust identification of cell lineages solely from scRNA-seq data. By leveraging gene expression memory, GEMLI unveils heritable patterns, cell fate decisions, and multicellular structures, shedding new light on processes like cancer invasiveness.

Introduction

What if you have to take a snapshot of a busy city during rush hour? Back and forth, vehicles move with their destinations. Can you tell which vehicles set off at the same time? It’s what scientists studying cell lineages have to face: the family trees of cells. In this case, traditional methods encompass marking individual cells, just like putting colored stickers on the vehicles. However, this can be technically difficult and restricts researchers to specific periods.

Notably, GEMLI (Gene Expression Memory-based Lineage Inference), a cutting-edge tool for such purposes, was invented by Dr. Almut Eisele et al., which is capable of tracking cellular ancestry only through monitoring their gene expression patterns – the particular sets of genes that are active in a cell at any given moment. By introducing the concept in their recent Nature Communications article, novel approaches negate extra experiments, allowing easy and keen study of lineage.

Continue Reading

The significance of comprehending gene expression regulation in tissue development is highlighted by the examination of disease progression through methods such as ChIP-seq and ATAC-seq. On the other hand, the growing volume of data poses computing difficulties, and the dearth of organized tools emphasizes the requirement for effective analysis platforms. Researchers from the Shanghai Institute of Nutrition and Health created the scalable cloud-based Epigenomic Analysis Platform (EAP, https://www.biosino.org/epigenetics) to solve these problems and analyze large-scale ChIP/ATAC-seq data sets effectively. 

In order to extract biologically significant insights from heterogeneous datasets and automatically produce publication-ready figures and tabular results, EAP uses sophisticated computational algorithms. This allows for thorough epigenomic analysis and data mining in fields like cancer subtyping and therapeutic target discovery.

Understanding the regulation of gene expression in tissue development and disease progression is largely dependent on epigenome profiling. Cutting-edge deep sequencing methods like ATAC-seq and ChIP-seq have made it possible to analyze epigenetic variation in disease cohorts and developmental cells, offering important new insights into the processes governing gene expression and the course of disease. However, to explore these massive datasets, researchers need greater computing power as data collection grows in scope. 

Continue Reading

Picture a world where computers can not only translate languages but also decipher biology’s convoluted language. This is the exciting frontier of Large Language Models (LLMs) that could transform our knowledge of genes and cells, which are the foundation of all life forms. Researchers from the Center, Chinese Academy of Sciences, China, explore this intriguing crossroad. Genes, which are passed down from one generation to another, hold the truths of our being. Cells, the tiny factories that keep us alive, execute these instructions coded in genes. Decoding how genes and cells interact helps to unravel health complications, diseases, and even mysteries regarding evolution.

Traditionally, scientists have used gene sequencing to study these intricate associations. But LLMs present an alternative way forward with immense promise. These models are trained using huge volumes of text data, enabling them to understand complicated patterns as well as mappings between them. Perhaps scientists can get a breakthrough by passing such datasets through LLMs.

Continue Reading

Are you aware that plenty of operational AI applications have emerged in the early stages of small-molecule drug discovery? Do you recognize that those packages regularly require a thorough grasp of software programs and hardware, as well as a specialization in an optimistic effort, inflicting boundaries for regular users, in fact, normalization and switching among them? Researchers at Zhejiang Shuren University and Hangzhou SanOmics Information Technology created MolProphet, a user-friendly software with an AI-based total interface to cope with this impediment. MolProphet offers a complete answer for small molecule drug development, together with AI-pushed target pocket prediction, hit discovery, lead optimization, compound concentration, and powerful analytics tools. Its design seeks to simplify the process, permitting novice users to navigate through the numerous tiers of drug research.

Continue Reading

The synergy among the Nuffield Department of Medicine and the Department of Statistics at the University of Oxford represents a blend of medical expertise with statistical rigor – “The MolSnapper,” built using deep learning, promptly exploring intricate molecular specificity for proteins in 3D conformations. Unlike previous approaches that rely on costly trial and error, Conditioning Diffusion uses deep learning to navigate the maze of molecular interactions with unprecedented efficiency to provide novel medical research solutions.

Their research bridges the gap by making it less complicated to evolve patterns learned from large molecular datasets for pocket-binding applications. MolSnapper employs a controlled and environmentally friendly producing approach to seamlessly combine 3-D structural insights with expert knowledge, revolutionizing the ligand era in pharmaceutical applications. MolSnapper enables researchers to generate molecules from large datasets by providing personalized instruction and enhanced selection. This differs from conventional structure-based drug design approaches, which are stereotypically proficient on restricted protein molecular data.

Precision is imperative in drug design and discovery. Therefore, MolSnapper authenticates this concept by serving to deliver personalized descriptions specific to the investigator’s requirements. With the potential for drug discovery employing deep learning, the path has been opened for personalized medicine intensive on molecular markers unique to particular diseases.

Continue Reading

Proteins are life’s molecular workhorses and play crucial roles in every biological process. Understandably, their complex shapes, also known as conformations, are the key that unlocks their secrets to new drug development. Proteins are not still statues but rather change into different conformations like ninjas to play different functions. This movement of protein has been a major challenge during drug discovery because traditional methods usually capture a single snapshot of a protein.

A glimmer of hope comes from a recent study published in Nature Communications by a team at Brown University. Their investigation introduces a brand new method that employs artificial intelligence (AI) to predict several shapes that proteins can adopt. If it is upheld, this advancement could alter our perception of protein functioning and lead to better medications.

As stated by the media release on the Brown University website, protein structures were conventionally identified using X-ray crystallography, among other techniques. Despite their value, however, these methods give a static image, just like a still photo of protein in one pose. Nonetheless, proteins are dynamic entities that change their shapes continuously to perform their function. This dynamic nature is important for understanding how drugs interact with proteins. A protein could be likened to a lock while the drug is seen as the key; if the protein changes shape, then the key may not fit any longer, and hence, this renders it useless for what it was meant for.

Continue Reading

Can you imagine that you can design proteins that have a particular function, and they are like small robots working inside our cells? This may be the reality pretty soon. Researchers from the University of California and Yale University came up with an innovative way of creating transmembrane proteins from scratch, thus leading to targeted treatment and a better understanding of cell biology. However, before we dive into this thrilling research, let us first look at what proteins and membranes are.

Picture your cells as something like busy factories. Proteins make sure these little factories keep running smoothly the entire time. They exist in a multitude of forms and sizes, each having its own work. Some function as enzymes, which are catalysts for chemical reactions, while others transport molecules across membranes, which enclose every cell as protective barriers.

There is a special kind of protein called Transmembrane proteins; it has a unique structure. It has got a hydrophobic section embedded within the cellular membrane, just like the screw goes into a cork bottle stopper. This part is known as transmembrane domain (TM domain). Depending on its purpose, the other part of the protein can stay in or outside the cell. These domains act as channels for molecule exchange between cells.

Continue Reading

Single-cell RNA sequencing (scRNA-seq), a rapidly developing field of RNA sequencing of single cells, has recently made a breakthrough that could revolutionize cell type annotation. Using GPT-4’s powerful language model, Columbia University Mailman School of Public Health and Duke University School of Medicine researchers have demonstrated that cell types can be accurately annotated based on marker gene information, greatly reducing the effort and expertise required for this critical step. The researchers additionally developed an R programming suite named GPTCelltype tailored for the automated annotation of cell types by GPT-4.

Although cell type annotation is a fundamental step in scRNA-seq analysis, it has traditionally been a long and laborious process. This analysis is typically done by a human expert by comparing highly expressed genes with canonical cell type markers. The task requires a lot of knowledge and experience.

Despite the development of automated cell-type annotation methods, manual annotation based on marker genes remains widely used, raising the need for more efficient and accurate solutions.

Continue Reading

AI is being used to find novel antibiotics for bacteria resistant to pandemics; however, the currently used techniques have serious drawbacks. While generative models design molecules—which are difficult to synthesize—property prediction models, which assess molecules individually, have difficulty scaling to enormous chemical spaces. The generative model SyntheMol, which creates readily synthesizable chemicals from a chemical space of 30 billion molecules, is presented here by researchers from Stanford University and McMaster University. Utilizing SyntheMol, compounds that inhibit the bacterial pathogen Acinetobacter baumannii were created. Six structurally unique compounds with strong action against Acinetobacter baumannii were among the 58 produced molecules that were synthesized and experimentally confirmed.

In modern medicine, the global spread of antibiotic resistance determinants poses a serious concern, with a projected 4.95 million deaths attributed to drug-resistant illnesses in 2019. Acinetobacter baumannii, a Gram-negative bacterium, is considered a critical priority by the World Health Organisation. Promising medication candidates, including antibiotics, can be quickly and precisely identified using artificial intelligence (AI) techniques. 

Although property prediction AI algorithms are time-consuming for vast chemical spaces, they are capable of evaluating chemical libraries to determine compounds with desirable properties. In contrast, generative models construct molecules from the ground up by fusing together smaller molecules with desired characteristics. This approach allows for the direct design of potential molecules without the need for a lengthy review process involving numerous compounds.

Continue Reading

The inability to logically design new antibodies that bind a particular epitope on a target persists despite the crucial role that antibodies play in contemporary medicine. Instead, time-consuming animal immunization procedures or library screening techniques are currently used in antibody development. Here, scientists from the University of Washington show that user-specified epitopes can be bound by de novo antibody variable heavy chains (VHHs) created by a refined RFdiffusion network. Researchers have confirmed binders to four disease-relevant epitopes through experiments, and the overall binding pose and CDR loop configuration of a proposed VHH bound to influenza hemagglutinin in the cryo-EM structure are almost the same as in the design model. 

Protein therapies, of which antibodies are the predominant class, now have over 160 licenses worldwide; in the next five years, the industry is projected to reach $445 billion. Therapeutic antibody development frequently depends on animal immunization or antibody library screening despite the great interest in these antibodies from the pharmaceutical industry. In addition to being time-consuming and difficult, these techniques may not provide antibodies that bind with the therapeutically important epitope. 

Several techniques have been used in the computational design of antibodies, such as utilizing the Rosetta sequence design, sampling different native CDR loops, and grafting residues onto pre-existing structures. De novo design of structurally correct antibodies has been difficult to achieve, though. Designing binding proteins with RFdiffusion has made it possible to create a wide variety of binders that are naturally shaped to complement the user-specified epitope. However, RFdiffusion is unable to design antibodies de novo.

Continue Reading

In the vast and intricate realm of metagenomics, where microbial communities unveil a wealth of biomedical knowledge, a groundbreaking development has emerged. Scientists from Pennsylvania State University introduced MetagenomicKG, a new knowledge graph built to take the exploration of metagenomic data to unprecedented heights. This revolutionary resource has the potential to change how researchers address the complexity of microbial ecosystems and provides a comprehensive and interconnected framework tailored to the unique needs of this emerging field.

The large amount and diversity of genomic content in microbial communities make metagenomics an affluent area of ​​biomedical knowledge. However, traversing these complex societies and their vast unknowns often depends on a variety of reference libraries, each with a specific analytical purpose. From the Genome Taxonomy Database (GTDB) and the Kyoto Encyclopedia of Genes and Genomes (KEGG) to the Bacteria and Viruses Bioinformatics Resource Center (BV-BRC), these repositories are essential for the genetic and functional annotation of microbial communities.

Despite their valuable contributions, inconsistent nomenclature or identifiers between these repositories present challenges for effective integration, representation, and use. Enter the Knowledge Graph, a powerful solution that organizes biological entities and their relationships into a coherent network, revealing hidden patterns and enriching our biological understanding with deeper insights.

Continue Reading