Citations

Abstract With the rapid advancements in large language model technology and the emergence of bioinformatics‐specific language models (BioLMs), there is a growing need for a comprehensive analysis of the current landscape, computational characteristics, and diverse applications. This survey aims to address this need by providing a thorough review of BioLMs, focusing on their evolution, classification, and distinguishing features, alongside a detailed examination of training methodologies, datasets, and evaluation frameworks. We explore the wide‐ranging applications of BioLMs in critical areas such as disease diagnosis, drug discovery, and vaccine development, highlighting their impact and transformative potential in bioinformatics. We identify key challenges and limitations inherent in BioLMs, including data privacy and security concerns, interpretability issues, biases in training data and model outputs, and domain adaptation complexities. Finally, we highlight emerging trends and future directions, offering valuable insights to guide researchers and clinicians toward advancing BioLMs for increasingly sophisticated biological and clinical applications.

Abstract Aging increases the vulnerability of kidneys to injury and impairs their regenerative capacity. SIRT3 expression declines with aging and is associated with multiple age‐related pathologies. The expression profile and functional role of SIRT3 in renal aging remain unclear. Here, SIRT3 expression in aging kidneys is assessed and analyzed for its promoter methylation patterns using methylation‐specific PCR (MSP). It is found that aging exacerbates UUO‐induced renal fibrosis, associated with downregulated SIRT3 expression. Mechanistically, age‐related SIRT3 downregulation is mediated by hypermethylation of its promoter region. SIRT3 knockout exacerbated renal fibrosis in young mice subjected to UUO, whereas SIRT3 overexpression attenuated fibrosis in aged UUO mice. Integration of RNA‐seq and immunoprecipitation‐mass spectrometry (IP‐MS) analyses revealed that SIRT3 deficiency leads to hyperacetylation of GSK3β at lysine 15 (K15). This K15 hyperacetylation inhibited GSK3β activity, consequently stabilizing its substrate β‐catenin. Furthermore, self‐assembled PEG‐PCL‐PEG micelles are designed and synthesized to encapsulate hydrophobic honokiol (HKL). These micelles significantly enhanced the aqueous solubility and oral bioavailability of free HKL, maintained stable blood concentrations, and ultimately improved its anti‐fibrotic efficacy. These findings propose novel therapeutic strategies for managing renal fibrosis in the aging population and provide a foundation for developing new drugs and combination therapies.

Abstract RATIONALE: Alveolar Type 2 (AT2) cells play a critical role in maintaining alveolar function and facilitating repair processes in the lung. Despite recent advancement on transcriptomic mapping of cell identities and AT fate selection, the mechanisms governing AT2 fate plasticity remain elusive. Notably, in AT2-derived pulmonary diseases such as COPD and IPF, disruptions in AT2 differentiation are observed, suggesting alterations may exist in the regulation of fate plasticity. Therefore, a refined model describing the regulation of AT2 fate plasticity is needed, along with an assessment of the model's conservation across murine and human AT2s. METHODS: We performed scRNA-sequencing, ATAC-seq, and bioinformatic analysis on samples across murine lung development and validating findings by immunostaining of embryonic lungs. We conducted organotypic cultures as well as genetically engineered mouse model experiments to assess changes in AT2 fate plasticity in. Further, we established a Mayo Clinic patient lung cell biorepository from patient tissue derived AT2s. In parallel, we performed long read scRNA-sequencing for isoform discovery, organotypic culture and immunostaining in human AT2 organoids, to characterize the presence of similar regulatory mechanisms. RESULTS: Throughout the late embryonic and early perinatal period, we found a window of fate plasticity exists in AT2s whereby they retain the capacity to give rise to AT1s, and that its being closed over time by the bZIP transcription factor C/EBPα, as its genetic deletion in AT2s results in regaining fate plasticity. By ATAC-seq and ChIP-seq, we found C/EBPα binds closely to the Dlk1 locus, a Notch signaling regulator implicated in AT2-to-AT1 conversion that is upregulated following Cebpa deletion. Following adult lung injury, we found C/EBPα downregulation is required for re-accessing AT2 fate plasticity and is likely mediated by the dominant negative C/EBP family member CHOP. Finally, we observe Cebpa loss also activates a “defender” AT2 state enriched in interferon stimulated antipathogenic gene expression that is distinct from its reparative state and propose AT2s can toggle between these states following infection to robustly protect and repair alveoli.Further, we validated the presence of C/EBPα and absence of Dlk1 in human AT2s isolated from healthy adult patients, which persists following in vitro expansion. We conclude that AT2 cells window of fate plasticity is dictated by a molecular circuit involving C/EBPα suppressing and Dlk1. Future works aim to remove C/EBPα in human AT2s to reopen their fate plasticity for cell therapy.

Abstract Background Efficient and accurate delineation of organs at risk (OARs) is a critical procedure for treatment planning and dose evaluation. Deep learning‐based auto‐segmentation of OARs has shown promising results and is increasingly being used in radiation therapy. However, existing deep learning‐based auto‐segmentation approaches face two challenges in clinical practice: generalizability and human‐AI interaction. A generalizable and promptable auto‐segmentation model, which segments OARs of multiple disease sites simultaneously and supports on‐the‐fly human‐AI interaction, can significantly enhance the efficiency of radiation therapy treatment planning. Purpose Meta's segment anything model (SAM) was proposed as a generalizable and promptable model for next‐generation natural image segmentation. We further evaluated the performance of SAM in radiotherapy segmentation. Methods Computed tomography (CT) images of clinical cases from four disease sites at our institute were collected: prostate, lung, gastrointestinal, and head & neck. For each case, we selected the OARs important in radiotherapy treatment planning. We then compared both the Dice coefficients and Jaccard indices derived from three distinct methods: manual delineation (ground truth), automatic segmentation using SAM's ’segment anything’ mode, and automatic segmentation using SAM's ‘box prompt’ mode that implements manual interaction via live prompts during segmentation. Results Our results indicate that SAM's segment anything mode can achieve clinically acceptable segmentation results in most OARs with Dice scores higher than 0.7. SAM's box prompt mode further improves Dice scores by 0.1∼0.5. Similar results were observed for Jaccard indices. The results show that SAM performs better for prostate and lung, but worse for gastrointestinal and head & neck. When considering the size of organs and the distinctiveness of their boundaries, SAM shows better performance for large organs with distinct boundaries, such as lung and liver, and worse for smaller organs with less distinct boundaries, like parotid and cochlea. Conclusions Our results demonstrate SAM's robust generalizability with consistent accuracy in automatic segmentation for radiotherapy. Furthermore, the advanced box‐prompt method enables the users to augment auto‐segmentation interactively and dynamically, leading to patient‐specific auto‐segmentation in radiation therapy. SAM's generalizability across different disease sites and different modalities makes it feasible to develop a generic auto‐segmentation model in radiotherapy.

Abstract SETDB1 is a key regulator of lineage-specific genes and endogenous retroviral elements (ERVs) through its deposition of repressive H3K9me3 mark. Apart from its H3K9me3 regulatory role, SETDB1 has seldom been studied in terms of its other potential regulatory roles. To investigate this, a genomic survey of SETDB1 binding in mouse embryonic stem cells across multiple libraries was conducted, leading to the unexpected discovery of regions bereft of common repressive histone marks (H3K9me3, H3K27me3). These regions were enriched with the CTCF motif that is often associated with the topological regulator Cohesin. Further profiling of these non-H3K9me3 regions led to the discovery of a cluster of non-repeat loci that were co-bound by SETDB1 and Cohesin. These regions, which we named DiSCs (domains involving SETDB1 and Cohesin) were seen to be proximal to the gene promoters involved in embryonic stem cell pluripotency and lineage development. Importantly, it was found that SETDB1-Cohesin co-regulate target gene expression and genome topology at these DiSCs. Depletion of SETDB1 led to localized dysregulation of Cohesin binding thereby locally disrupting topological structures. Dysregulated gene expression trends revealed the importance of this cluster in ES cell maintenance as well as at gene ‘islands’ that drive differentiation to other lineages. The ‘unearthing’ of the DiSCs thus unravels a unique topological and transcriptional axis of control regulated chiefly by SETDB1.