Institute for Systems Biology

Efficient genome writing in mammalian cells requires robust methods for integrating large DNA payloads. The previously described method mammalian Switching Antibiotic resistance markers Progressively for Integration (mSwAP-In) enables iterative, biallelic genome rewriting in mammalian stem cells with DNA payloads exceeding 100 kb. However, the lack of standardized vectors and certain technical constraints have limited its broader adoption. Here we present an improved plasmid toolkit designed to streamline the implementation of mSwAP-In. The toolkit includes two core vectors. pLP-TK (pCTC174) is a landing-pad plasmid compatible with Golden Gate assembly of genomic homology arms and supports both mSwAP-In and the recombinase-mediated cassette exchange method Big-IN. mSwAP-In MC2v2 (pKBA135) is a versatile Big DNA assembly and delivery vector that supports Gibson-based assembly and incorporates positive, negative, and fluorescent selection markers, as well as a backbone counterselection cassette to minimize unwanted plasmid integration. The vector architecture also enables propagation in yeast and bacterial hosts, inducible plasmid copy-number amplification in standard E. coli strains, and CRISPR/Cas9-mediated payload release through preinstalled guide RNA target sites. We further characterize the FCU1/5-FC counterselection system in mouse embryonic stem cells and define conditions that minimize its bystander toxicity. Finally, we provide a set of Cas9-gRNA expression plasmids optimized for common mSwAP-In applications. Together, these reagents constitute a standardized and experimentally validated toolkit that simplifies large-scale genome writing using mSwAP-In.

The field of human and medical genomics is undergoing two major transformations (Figure 1). First, data density and dimensionality is increasing. Second, a combination of new and old analytical techniques-notably empowered by artificial intelligence (AI)-are enabling extraction of mechanistic insight and knowledge from these data. In this editorial we highlight a fraction of the many key papers recently published in this field, with a particular emphasis on those that were invited to our research topic with speculative insights based on literature review or research that reflected state-of-the-art methods, detailed recent developments, or highlighted pivotal accomplishments. A focus of this research topic was to showcase progress and pinpoint hurdles that need to be surmounted to push the boundaries of what is medically and technically feasible in genetics.Two years ago, at the launch of this research topic, we envisioned potential advances in the following categories: (1) novel genomic technologies and their clinical application; (2) genetic determinants of complex diseases; (3) ethical, legal, and social implications of genomic research; (4) personalized medicine and genomic tailoring; (5) advances in genomic sequencing techniques; and (6) integration of genomic data with AI and machine learning approaches. These fit into a broader arc of progress in our field. Human genomics fundamentally starts with the genome and ends with personalized medicine. Variant annotation remains the foundation of human genomics and must be built on careful observation of patients 1 reported on the scale of case reports 2 all the way up to population compilations. 3,4 AlphaMissense provides an example of advancing sophistication in variant annotation (https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2024.1487608/full). The insight is that variant annotation cannot end without consideration of complex genomic context. To advance mechanistic insights the field will increasingly incorporate epistasis, 5 synergies, population and ancestry context, 6 and whole genome interactions into this functional foundation. Viewing the genome as a web of interactions, and not as a linear construct is a necessary start. Pangenome representations (https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2025.1679660/full) illustrate an important step in this direction. Furthermore, The Peruvian Genome Project (https://www.frontiersin.org/journals/genetics/articles/10.3389/fgene.2025.1614021/full) highlights the importance of worldwide pan-population data generation coupled to FAIR data governance (https://www.go-fair.org/fair-principles/). Once a foundation of data is generated, it must be analyzed. The insight is that this analysis must be increasingly mechanistic and holistic, focusing on Self-critique of the field must address major questions:Are current omics data collections big enough? Dense data drives twenty-first century research. 10 Even though modern data sets may comprise data on thousands of variables (e.g., a combination of plasma proteomics and metabolomics), these may not be enough to truly illuminate the dynamic state of the entire human systems and its environment. A combination of increasing technological efficiency, brute force of scale, and careful hypothesis driven selection of assayed biological targets will be necessary to drive complete insights (e.g., shifting from left to right in Figure 1). Are we examining enough dimensions? Many analyses focus on genomics coupled to blood proteomics, metabolomics, & epigenetics. The microbiome is gaining increasing attention. For better systems understanding, we will need to look at multiple tissues and cells (and their compartments) simultaneously, and to measure the entire system dynamically over time. Larger, less biased, recruitment must occur in clinical trials and studies. Continuous monitoring devices and solutions that enable remote trials with massive population participation will drive the future.As analyses get increasingly complex, it becomes more important to carefully design controls, and to red-team (systematically testing for flaws or biases) the computational workflows. Control analyses limit the ways our complex workflows can fool us into conclusions not supported by the data. The safeguards provided by controls become critically important given the widespread availability of AI-and ML-tools that automate data handling without necessarily exercising judgment or understanding. We must use systems thinking and rational epistemological approaches such as Hill's criteria to arrive at medically useful translatable knowledge. 7Are we creating papers and disseminating data that will be part of large meta-analyses? As research becomes increasingly collaborative and cross-discipline, we must write manuscripts that will be accessible and understood by lay people, students, experts, and AI agents. Datasets must be publicly available, easing any barriers to high-throughput automated access, following the FAIR

Centromeres ensure accurate chromosome segregation, yet their DNA evolves rapidly across eukaryotes leaving the origins of new centromere architectures unclear1-4. The brewer's yeast Saccharomyces cerevisiae exemplifies this long-standing puzzle. Its centromeres shifted ancestrally from large, repeat-rich, epigenetically specified forms to the compact, genetically defined 'point' centromeres1,5. How this transition occurred has remained unresolved6. Here we identify evolutionarily related 'proto-point' centromeres that provide a resolution to the evolutionary origins of point centromeres. Proto-point centromeres contain a single centromeric nucleosome positioned over an AT-rich core, accompanied by relaxed organization and sequence variability of flanking cis-elements. In two species, these proto-point centromeres lie within retrotransposon-derived repeat clusters, linking ancestral repeat-rich centromeres to genetically encoded ones. Comparative and phylogenetic analyses indicate that proto-point and point centromeres evolved in an ancestor with retrotransposon-rich centromeres. These results identify long-terminal-repeat retrotransposons, specifically Ty5 sequences, as the genetic substrate for point-centromere evolution and provide a mechanistic route by which an epigenetic centromere can become genetically specified. More broadly, they show how selfish elements can be co-opted to perform essential chromosomal functions.

Prebiotic, probiotic, and combined (synbiotic) interventions often show variable outcomes across individuals, driven by complex interactions between introduced biotics, the endogenous microbiota, and the host diet. Predicting individual-specific success or failure of probiotic and prebiotic therapies remains a major challenge. Here, we leverage microbial community-scale metabolic models (MCMMs) to predictprobiotic engraftment and microbiota-mediated short-chain fatty acid (SCFA) production in response to probiotic and prebiotic interventions. Using data from two human clinical trial cohorts, testing a five-strain probiotic combined with the prebiotic inulin designed to improve metabolic health and an eight-strain probiotic designed to treat recurrent Clostridioides difficile infections, respectively, we show that MCMM-predicted engraftment largely agrees with measurements, achieving 75%-80% accuracy. Engraftment probabilities varied across taxa. MCMMs captured treatment-driven shifts in predicted SCFA production, and higher model-predicted growth rates of Akkermansia muciniphila were negatively associated with glucose area under the curve (AUC) in the first trial, providing clues about the mechanisms underlying treatment efficacy. Extending these models to a third human cohort undergoing a healthy diet and lifestyle intervention revealed substantial inter-individual variability in predicted responses to increasing dietary fiber, which were significantly associated with baseline-to-follow-up changes in cardiometabolic health markers. Finally, our simulation results suggested that personalized prebiotic selection may further enhance probiotic efficacy. Together, these findings demonstrate the potential of metabolic modeling to guide personalized microbiome-mediated interventions.

Precision medicine demands a shift from static, single-analyte diagnostics toward dynamic, systems-level understanding of health and disease. This review explores how the convergence of systems biology, multiomics, and artificial intelligence (AI) redefines biomarker discovery to drive early disease detection and personalized intervention. We highlight pioneering efforts that use longitudinal, multimodal data to map individual health trajectories and uncover early disease signals. Advances in AI, including machine learning and contextualization using knowledge graphs and digital twins, are accelerating clinical translation by enabling predictive, context-aware analyses. Real-world applications, including omics-informed diagnostics and digital health monitoring, demonstrate the potential of this approach to transform health care from reactive treatment to proactive wellness. These technologies also inform the development of targeted therapeutics that intervene earlier, personalize treatment, and potentially halt or reverse disease progression. We outline challenges, emerging solutions, and future directions that position AI-driven systems biology at the center of next-generation precision health.

Abstract Background Post Acute Sequelae of COVID-19 (PASC), also referred to as Long COVID, is an infection-associated chronic syndrome with heterogenous symptom profiles that occurs in a subset of people following SARS-CoV-2 infection. Despite proposed viral persistence mechanisms, no therapeutic benefit was observed in two randomized placebo-controlled trials of nirmatrelvir/ritonavir (NMV/r) in adults with Long COVID, including the Selective Trial of Paxlovid for PASC (STOP-PASC) and PAX LC. This systems immunology analysis aimed to characterize immune profiles of participants during clinical trial intervention, identify biomarkers associated with patient-reported outcomes, and investigate potential mechanisms underlying Long COVID. Methods We performed comprehensive immunological profiling of 152 STOP-PASC trial participants using plasma proteomics (Olink® Explore HT 5400 panel), autoantigen arrays, viral serology, and microclot assays at baseline, day 15, and week 10. We assessed associations between immune features and patient-reported outcomes. We also conducted meta-analysis of nine independent Long COVID proteomics cohorts (n=590 total samples) to identify conserved inflammatory signatures. Results NMV/r treatment at day 15 compared with baseline induced transient changes in plasma proteins that normalized by week 10, primarily impacting myeloid cell/monocyte, lysosome, and complement activation pathways. Cardiovascular symptoms were negatively associated with SARS-CoV-2 antibody levels at baseline. No widespread differences in autoantibody profiles, Epstein-Barr virus (EBV) reactivation, or microclotting were observed between STOP-PASC Long COVID participants, pre-pandemic controls, and individuals without Long COVID. Meta-analysis of publicly available Olink® data from Long COVID cohorts identified a conserved 60-protein Long COVID Signature (LCS) score revealing multi-compartment immune activation involving monocyte, neutrophil, and T/NK cell modules. Conclusions These findings advance our understanding of Long COVID immunology and may help direct future proteomic biomarker endpoints for Long COVID clinical trials.

BackgroundLipid biology is profoundly disrupted in AD, as evidenced by abundant genetic studies highlighting the roles of lipid metabolism, endosomal/lysosomal trafficking, bioenergetics, and other cellular and molecular alterations. Despite these observations, there is currently no unified hypothesis to explain the etiology of AD that reconciles lipid disruption with neuronal loss.MethodUsing baseline measurements from ADNI plasma lipidomics and CSF proteomics, extensive reconciliation of the literature, and methods related to systems engineering and QSP, we built a model of brain homeostasis centered around feedback loops and mass balance. We then apply our model to several large multiomic datasets, including those in ADNI and GNPC for validation.ResultWe identified multiomic/proteomic evidence of two convergent feedback loops driving AD pathophysiology: 1) Saturated Microglial Flux Capacity: Microglia become overwhelmed by neuronal and myelin debris, leading to impaired processing. 2) Impaired ANLS Signaling: Elevated astrocyte membrane cholesterol disrupts ANLS, compromising neuronal metabolic support.ConclusionNeurodegeneration in AD is driven by the brain's failure to process excess cholesterol from neuronal debris, resulting in widespread lipidomic, metabolic, and proteomic disruptions. Disruptions in lipid homeostasis occur upstream of both amyloid and tau pathology, as excess cholesterol accumulation enhances APP translocation to γ‐secretase, accelerating amyloid production, while also impairing astrocytic metabolic support, leading to ATP deficits, elevated intracellular Ca²⁺, and aberrant tau phosphorylation. This breakdown in cholesterol homeostasis offers a comprehensive explanation for the observed multi‐omic alterations in AD, positioning lipid dysregulation as a primary driver of disease progression. Our unified hypothesis enables us to leverage large omics data to learn new biology, and identify and evaluate novel therapeutic targets for both the prevention and treatment of AD.

When listening to speech in their native language, speakers use prior context to anticipate upcoming phonemes, words, and concepts, integrating information at the sublexical, lexical, and sentence level. While it has been suggested that late second language learners do not predict to the same extent as native listeners, adequately evaluating this claim requires measurement of predictions at these multiple levels of representation simultaneously in natural speech. We recorded magnetoencephalography (MEG) responses from native Mandarin and Sinhala speakers listening to continuous narrative English speech. We used multivariate temporal response function (mTRF) analysis to investigate whether second language listeners demonstrate the same markers of prediction in neural data as native English speakers listening to the same stimuli. We demonstrate that late second language listeners exhibit strikingly similar responses to native speakers in sensitivity to phoneme surprisal and entropy with respect to sublexical, lexical, and sentence-level context. The few small response differences we observed appear most likely to arise from specific properties of the native languages, rather than general differences between native and second-language listening. These results provide evidence that late second-language listeners indeed leverage prediction in similar ways as native listeners in understanding continuous speech. This suggests that multivariate analyses of neural data from naturalistic listening may be vital in carefully evaluating the differences and similarities in speech prediction across populations.