Structural maintenance of chromosomes (SMC) complexes are essential proteins found in genomes of all cellular organisms. Essential functions of these proteins, such as mitotic chromosome formation and sister chromatid cohesion, were discovered a long time ago. Recent advances in chromatin biology showed that SMC proteins are involved in many other genomic processes, acting as active motors extruding DNA, which leads to the formation of chromatin loops. Some loops formed by SMC proteins are highly cell type and developmental stage specific, such as SMC-mediated DNA loops required for VDJ recombination in B-cell progenitors, or dosage compensation in Caenorhabditis elegans and X-chromosome inactivation in mice. In this review, we focus on the extrusion-based mechanisms that are common for multiple cell types and species. We will first describe an anatomy of SMC complexes and their accessory proteins. Next, we provide biochemical details of the extrusion process. We follow this by the sections describing the role of SMC complexes in gene regulation, DNA repair, and chromatin topology.

A bstract: Recent advances in DNA language models have mainly come from building larger and more complex architectures, making it harder to understand the effect of changes to standard components such as the transformer layers widely used in NLP. In this work, we study whether and how a modernized BERT-style back-bone (ModernBERT) can be adapted to genomic sequence modeling to improve computational efficiency, training stability, and downstream performance. Under controlled experimental settings, we benchmark efficiency across a range of sequence lengths and evaluate downstream performance on the Nucleotide Transformer benchmark. The resulting model, ModernGENA, achieves a strong efficiency–quality trade-off and ranks among the top-performing models in our evaluation suite. To support reproducibility and provide a solid default reference point for future architectural work in genomics, we release the full implementation and configuration of ModernGENA as an open, reusable baseline, and make ModernGENA base and ModernGENA large publicly available through the DNA language models collection on Hugging Face .

This study presents a novel approach for mapping global chromatin interactions using S1 nuclease, a sequence-agnostic enzyme. We develop and outline a protocol that leverages S1 nuclease’s ability to effectively introduce breaks into both open and closed chromatin regions, allowing for comprehensive profiling of chromatin properties. Our S1 Hi-C method enables the preparation of high-quality Hi-C libraries, marking a significant advancement over previously established DNase I Hi-C protocols. Moreover, S1 nuclease’s capability to fragment chromatin to mono-nucleosomes suggests the potential for mapping the three-dimensional organization of the genome at high resolution. This methodology holds promise for an improved understanding of chromatin state-dependent activities and may facilitate the development of new genomic methods.

Inference of gene structure and location from genome sequences - known as de novo gene annotation - is a fundamental task in biological research. However, sequence grammar encoding gene structure is complex and poorly understood, often requiring costly transcriptomic data for accurate gene annotation. In this work, we benchmark current solutions and develop new methods of gene annotation. We show that pre-trained DNA language model (DNA LM) embeddings do not capture the features necessary for precise gene segmentation, and that task-specific fine-tuning remains essential. We comprehensively evaluate the impact of model architecture, training strategy, receptive field size, dataset composition, and data augmentations on gene segmentation performance. We revisit standard evaluation protocols, showing that commonly used per-token and per-sequence metrics fail to capture the challenges of real-world gene annotation. We introduce and theoretically justify new biologically grounded metrics, along with benchmarking datasets that better capture annotation quality. We show that fine-tuned DNA LMs outperform existing annotation tools, generalizing across species separated by hundreds of millions of years from those seen during training, and providing segmentation of previously intractable non-coding transcripts and untranslated regions of protein-coding genes. Our results thus provide a foundation for new biological applications centered on accurate gene annotation.

Loop extrusion by SMC complexes is a key mechanism underlying chromatin folding during both interphase and mitosis. Despite this shared mechanism, computational models of loop extrusion often rely on fundamentally different assumptions: interphase models typically use dynamic extruders with finite lifetimes, whereas mitotic models employ static extruders placed according to loop size distributions. In this work, we investigate whether these modeling paradigms are interchangeable or yield intrinsically incompatible results. Using publicly available Hi-C data from mitotic chicken cells, we systematically compare dynamic and static loop extrusion models implemented in the Polychrom framework. We evaluate how key parameters such as the extruder lifetime, extrusion velocity, and spatial noise affect the simulated contact probability curves P(s) and loop size distributions. Our results reveal that while both model types can be tuned to approximate the general shape of P(s), they produce distinct internal structures and divergent relationships between loop size and contact decay. We also show that increased extruder lifetimes lead to excessive nested loop formation, which alters both loop statistics and P(s) derivatives. Introducing spatial exclusion constraints between extruders partially restores consistency with static models. These findings highlight that differences in extruder behavior and polymer noise levels can significantly impact chromatin model outcomes and must be carefully accounted for when interpreting or comparing simulation results across biological conditions. Author summary: Chromatin organization plays a crucial role in gene regulation and cellular function, yet our understanding of its three-dimensional structure relies heavily on computational modeling and the interpretation of complex experimental data. In this study, we use coarse-grained modeling approaches to simulate chromatin folding and systematically investigate how different analysis metrics and data processing methods influence the conclusions drawn from such models. By comparing widely used metrics and exploring the effects of normalization and noise, we highlight potential pitfalls and biases that can arise in chromatin modeling studies. Our findings provide practical recommendations for researchers in the field, aiming to improve the robustness and reproducibility of computational analyses of chromatin architecture. This work will help guide future studies toward more reliable interpretations of chromatin structure and its biological implications.