Guangrong Qin

Logical gene regulatory network (GRN) models provide interpretable, mechanistic representations of cellular regulation and are widely used in systems biology. However, most existing models remain incomplete, context-specific, and difficult to extend to comprehensive GRNs, limiting their broader applicability to tasks such as drug-response prediction and precision medicine. Here, add some ideas about virtual twins: With AML patients, we can use comprehensive GRNs to predict drug response. We present KGBN (Knowledge Graph–augmented Boolean Network modeling), a computational workflow for systematically augmenting logical GRN models. KGBN incorporates regulatory interactions derived from curated knowledge graphs as alternative logical rules while preserving the validated structure of existing models. Rule probabilities are optimized against experimental data to represent regulatory uncertainty and achieve data-driven calibration. Applying KGBN to acute myeloid leukemia, we show that extending an existing GRN with drug-target pathways and training against ex vivo drug-response data yields mutation-specific models that recapitulate known therapeutic sensitivities and signaling dependencies, demonstrating the utility of KGBN for interpretable, context-aware GRN modeling. (Should be predict)

The B‐cell lymphoma 2 (BCL‐2) inhibitor venetoclax (VEN) in combination with hypomethylating agents (HMAs) has improved treatment outcomes for acute myeloid leukemia (AML) patients unfit for intensive chemotherapy and is increasingly used in the relapsed/refractory setting. However, primary resistance remains a significant challenge, affecting 20%–35% of treatment‐naïve and around 50% of previously treated AML patients. To investigate the mechanisms driving primary resistance to VEN–HMA therapy, we analyzed genetic, transcriptomic, BCL‐2 family protein expression, and ex vivo drug sensitivity data from 101 AML patients and correlated these profiles with clinical outcomes to VEN–HMA. Our study found that blasts from refractory patients exhibit an elevated BCL‐XL/BCL‐2 protein expression ratio, an immature CD34 ⁺ CD38 ⁻ phenotype, and frequent TP53 mutations. Consistent with the high ratio of BCL‐XL/BCL‐2, resistant samples showed increased ex vivo sensitivity to the dual BCL‐2/BCL‐XL inhibitor navitoclax. In addition, SMAC mimetics were effective in refractory blasts, which correlated with high TNF gene expression in these cells. Ex vivo treatment with the combination of navitoclax and SMAC mimetics further enhanced the eradication of VEN–HMA refractory blasts, although toxicity was also observed in healthy CD34 ⁺ cells. In conclusion, our integrative analysis identifies molecular signatures associated with primary VEN–HMA resistance and highlights BCL‐2/BCL‐XL inhibition and SMAC mimetics as therapeutic strategies to target resistance.

Nucleophosmin-1 ( NPM1 ) mutations define a major molecular subtype of acute myeloid leukemia (AML) and is generally associated with favorable prognosis. However, the impact of myelodysplasia-associated mutations (MDSm+) on patient outcomes within this subgroup remains uncertain. We retrospectively analyzed 271 NPM1 -mutated AML patients from three independent cohorts (SWOG, Fred Hutch, and Beat AML) to assess the prognostic significance of MDSm+ and its interaction with age. MDSm+ occurred in 17% of cases, most commonly involving SRSF2 and SF3B1 . Although MDSm+ was associated with inferior overall survival compared to MDSm-in ELN2022 favorable-risk patients (HR 2.0, p =0.008), this effect was largely driven by worse outcomes in older patients ( _≥ 65 years) as older ELN22 favorable-risk patients had poor OS regardless of presence of MDSm+ compared to younger patients. After stratification of patients by age, there was not a significant difference between MDSm+ and MDSm-in either younger patients (HR 0.99, p=0.98) or older patients (HR 1.42, p =0.33). These findings indicate that MDSm+ in NPM1 + AML is not independently associated with adverse risk after adjusting for age and highlight the need for age-adjusted AML risk models.

Therapeutic toxicity, which can be life-threatening, presents a major challenge in treating patients with acute myeloid leukemia (AML). Medical digital twins, which are virtual representations of patient disease, have the potential to forecast disease progression and simulate potential treatments. Using neutrophil counts and blast percentages, we developed mechanistic models to predict toxicity (neutropenia) in AML patients receiving combination venetoclax and azacitidine treatment. We identified a best-fitting model, though patient-specific accuracy was highly variable. To address this variability, we investigated subsets of patients based on their accordance with model assumptions, and were able to identify features predictive of model fit. In addition, we found that continuous updating over time improves model accuracy. The model evaluated in this study could be further validated in a larger clinical setting and may support a digital twin for decision making in forecasting therapeutic toxicity of venetoclax and azacitidine treatment.

Summary/ Glioblastoma stem cell (GSC) cultures are initiated from glioblastoma (GBM) surgical resection tissue. They can capture and propagate key GBM primary tumor molecular and cellular features. We have deeply characterized four isocitrate dehydrogenase (IDH)-expressing (or IDH+) GSC cultures from unrelated adults to serve as cellular models for the majority of adult primary GBM. We demonstrate that GSC cultures can be continuously propagated in defined, serum-free media and 5% oxygen without requiring specialized growth substrates; have well-defined genomic and mtDNA variants and gene/protein expression profiles; and highly reproducible dose-survival curves when treated with the GBM standard-of-care therapies of ionizing radiation (IR) and temozolomide (TMZ). We also illustrate how expressed lentiviral barcodes, mtDNA variants and single cell gene expression profiling can be used to define and track cellular heterogeneity over 40 days after IR treatment. These well-characterized IDH+ GSC cultures can support many high throughput in vitro assay formats, including xenograft, organoid and other GBM disease modeling protocols. They should prove a useful resource to better understand GBM biology, and to identify new and more effective GBM therapies and treatment regimens. Highlights: Glioblastoma (GBM)-derived IDH-expressing Glioma Stem Cell (GSC) cultures can capture and propagate GBM genomic variants, gene and protein expression programs. Mitochondrial DNA (mtDNA) variants identified by high accuracy Duplex DNA sequencing are abundant and sub-clonally organized in GSC cultures. GSC cultures have highly reproducible dose-survival curves for ionizing radiation and temozolomide, the GBM standard-of-care therapies. GSC culture cellular heterogeneity can be captured, characterized and tracked by using expressed lentiviral barcodes, mtDNA variants and scRNA sequencing. eTOC blurb: IDH-expressing glioma stem cell cultures (GSCs) are experimentally tractable and versatile glioblastoma (GBM) cellular disease models. Deeply characterized GSC cultures can enable new work to understand GBM biology, and help identify new and more effective GBM therapies and treatment regimens. Graphical

Background: Gene regulatory network (GRN) models provide mechanistic understanding of genetic interactions that regulate gene expression and, consequently, influence cellular behavior. Dysregulated gene expression plays a critical role in disease progression and treatment response, making GRN models a promising tool for precision medicine. While researchers have built many models to describe specific subsets of gene interactions, more comprehensive models that cover a broader range of genes are challenging to build. This necessitates the development of approaches for improving the models through model merging. Results: We present LM-Merger, a workflow for semi-automatically merging logical GRN models. The workflow consists of five main steps: (a) model identification, (b) model standardization and annotation, (c) model verification, (d) model merging, and (e) model evaluation. We demonstrate the feasibility and benefit of this workflow with two pairs of published models pertaining to acute myeloid leukemia (AML). The integrated models were able to retain the predictive accuracy of the original models, while expanding coverage of the biological system. Notably, when applied to a new dataset, the integrated models outperformed the individual models in predicting patient response. Conclusions: This study highlights the potential of logical model merging to advance systems biology research and our understanding of complex diseases. By enabling the construction of more comprehensive models, LM-Merger facilitates deeper insights into disease mechanisms and enhances predictive modeling for precision medicine applications. Clinical trial number: Not applicable.

The growing availability of biomedical data offers vast potential to improve human health, but the complexity and lack of integration of these datasets often limit their utility. To address this, the Biomedical Data Translator Consortium has developed an open‐source knowledge graph–based system—Translator—designed to integrate, harmonize, and make inferences over diverse biomedical data sources. We announce here Translator's initial public release and provide an overview of its architecture, standards, user interface, and core features. Translator employs a scalable, federated, knowledge graph framework for the integration of clinical, genomic, pharmacological, and other biomedical knowledge sources, enabling query retrieval, inference, and hypothesis generation. Translator's user interface is designed to support the exploration of knowledge relationships and the generation of insights, without requiring deep technical expertise and gradually revealing more detailed evidence, provenance, and confidence information, as needed by a given user. To demonstrate Translator's application and impact, we highlight features of the user interface in the context of three real‐world use cases: suggesting potential therapeutics for patients with rare disease; explaining the mechanism of action of a pipeline drug; and screening and validating drug candidates in a model organism. We discuss strengths and limitations of reasoning within a largely federated system and the need for rich concept modeling and deep provenance tracking. Finally, we outline future directions for enhancing Translator's functionality and expanding its data sources. Translator represents a significant step forward in making complex biomedical knowledge more accessible and actionable, aiming to accelerate translational research and improve patient care.

As large clinical and multiomics datasets and knowledge resources accumulate, they need to be transformed into computable and actionable information to support automated reasoning. These datasets range from laboratory experiment results to electronic health records (EHRs). Barriers to accessibility and sharing of such datasets include diversity of content, size and privacy. Effective transformation of data into information requires harmonization of stakeholder goals, implementation, enforcement of standards regarding quality and completeness, and availability of resources for maintenance and updates. Systems such as the Biomedical Data Translator leverage knowledge graphs (KGs), structured and machine learning readable knowledge representation, to encode knowledge extracted through inference. We focus here on the transformation of data from multiomics datasets and EHRs into compact knowledge, represented in a KG data structure. We demonstrate this data transformation in the context of the Translator ecosystem, including clinical trials, drug approvals, cancer, wellness, and EHR data. These transformations preserve individual privacy. We provide access to the five resulting KGs through the Translator framework. We show examples of biomedical research questions supported by our KGs, and discuss issues arising from extracting biomedical knowledge from multiomics data.

Gene regulatory network (GRNs) models provide mechanistic understanding of the gene regulations and interactions that control various aspects of cellular behaviors. While researchers have constructed GRNs to model specific sets of gene regulations or interactions, little work has been made to integrate or merge these models into larger, more comprehensive ones that could encompass more genes, and improve the accuracy of predicting biological processes. Here, we present a workflow for merging logical GRN models, which requires sequential steps including model standardization, reproducing, merging and evaluations, and demonstrate its application in acute myeloid leukemia (AML) study. We demonstrate the feasibility and benefits of model merging by integrating two pairs of published models. Our integrated models were able to retain similar accuracy of the original publications, while increasing the coverage and explainability of the biological system. This approach highlights the integration of logical models in advancing system biology and enhancing the understanding of complex diseases. Author summary In our study, we tackle the challenges of integrating gene regulatory network (GRN) models to enhance our understanding of complex biological systems. GRNs are essential tools for understanding how genes regulate various cellular behaviors, but individual models often focus on specific sets of genes or interactions. We present a novel workflow that merges these individual logical GRN models into more comprehensive ones, providing a broader view of gene regulation. We applied this workflow to Acute Myeloid Leukemia (AML), a highly aggressive form of blood cancer. AML is challenging to treat due to its genetic complexity and the frequent occurrence of treatment-resistant mutations. Our integrated models retain the accuracy of the original models while offering improved coverage of the biological processes. This approach offers valuable insights into the disease’s underlying mechanisms through a combination of models that describe different aspects of AML. We envision that the proposed workflow will improve predictions, generate deeper insights, and improve our understanding and treatment of complex diseases like AML.