AI for Science.
Systems medicine.
Translational biomedicine.
Across biological scales
Explainable AI, multimodal clinical modeling, survival and longitudinal analysis, and decision-support frameworks for risk stratification and translational discovery across major chronic diseases.
Integration of transcriptomic, proteomic, metabolomic, and single-cell data to resolve disease heterogeneity, prioritize mechanistic drivers, and advance biomarker translation.
Plaque vulnerability, vascular microenvironment, vascular smooth muscle cell plasticity, single-cell states, and cross-scale mechanisms of vascular remodeling.
Systems-level investigation of circadian disruption, chronobiology, and vascular inflammation, linking temporal dysregulation to molecular states, disease progression, and clinical phenotypes.
Selected publications from the complete author-controlled record. Follow each research page to its evidence, scope and original sources.
Science Letter proposing current, linked public records of human protocols and major outcomes in individualized genetic therapy research; reporting proposals, not a new clinical trial.
Explainable survival-risk assessment for advanced heart failure with renal dysfunction, integrating multicenter clinical data, survival machine learning and external validation.
Systems-biology framework integrating multicohort and single-cell transcriptomics with circadian perturbation and CRC-cell experiments to quantify circadian misalignment and prioritize SYNE1.
A machine-learning and transcriptomic framework for quantifying atherosclerotic plaque vulnerability using bulk and single-cell data.
An integrative single-cell and bulk-transcriptomic study of vascular smooth muscle cell fate decisions and molecular heterogeneity in atherosclerosis.
Cross-disease transcriptomic systems-biology analysis identifying shared inflammatory signatures between independently profiled COVID-19 and heart-failure datasets and prioritizing TLR4, TLR2, STAT3, IL1B and CXCL8 through network and external expression analyses.
Mechanistic atherosclerosis study combining human plaque observations, mouse genetics and macrophage experiments to link KIF13B with MERTK-dependent efferocytosis and lipid-independent myeloid protection in experimental models.
Narrative review integrating circadian clock biology, immune-cell trafficking and vascular inflammation across cardiovascular disease, while distinguishing strong mechanistic and preclinical chronotherapy rationale from more heterogeneous and unresolved clinical timing evidence.
RAP-AIscore is an interpretable postoperative overall-survival model developed in a retrospective two-center study of 506 patients undergoing robot-assisted proctectomy, with internal validation and independent retrospective SYEMH evaluation.
A mechanistic translational preclinical study combining associative human carotid-plaque evidence with mouse atherosclerosis genetics, pooled single-cell analysis, primary human aortic smooth-muscle-cell perturbation, biochemical rescue and preclinical pharmacology to investigate how KIF13B regulates VSMC state through KCTD10-dependent KLF4 turnover.
A retrospective public-cohort transcriptomic study identified three circadian-related LUAD clusters and examined outcome, pathway, genomic and immune associations. External template assessment and computational treatment hypotheses are separated from prospective clinical utility.
A BODIPY-peptide redox probe was evaluated in solution and a small atherosclerosis mouse experiment, using reversible PA intensity changes at 740 nm and accompanying histology. The evidence supports the tested preclinical imaging approach, without human diagnostic, treatment or safety validation.
Stage-associated DNB analysis, TCPSLevel, a compact subtype classifier and ASPH follow-up are presented as distinct evidence layers.
Human plasma exosomal miRNA profiling in DCM with chronic heart failure identified an article-reported 92-feature differential set, computational target networks and external qRT-PCR assessment of 10 hub miRNAs in 10 patients and 6 controls.
A preclinical multimodal study evaluates MCC950 in B16F10-bearing mouse models of ICI-associated cardiac injury, combining cardiac and tumor phenotyping with immune profiling and supportive public human PBMC association data.
A small prospective single-arm study reported apatinib tumor-response, survival and toxicity observations. Enrollment, safety and efficacy denominators differ; the activity signal and descriptive safety findings do not establish comparative superiority or a general treatment recommendation.
A retrospective human-expression study developed a UC-versus-control classifier, evaluated it across external bulk datasets and added a separate UCRScore interpretation layer. Classification is not future-onset prediction; the reported single-cell extension has an unresolved accession-provenance mismatch.
A preclinical study evaluated peptide-functionalized miR-185-5p-inhibitor liposomes in DOX-injured cells and mice, with formulation, biodistribution, cardiac phenotyping and cell-death-associated readouts. No human therapeutic efficacy was established.
HCCcrds-related subtype analysis separates expression, retrospective outcome associations and external template assessment.
Exploratory plasma exosomal circRNA profiling in DCM-HF reports 49 differential circRNAs and a small external qRT-PCR evaluation. The evidence page explicitly distinguishes the Methods microarray platform from sequencing terminology elsewhere in the article.
Integrated human single-cell and bulk transcriptomic analyses describe AAA-associated T-cell state heterogeneity, derive TIRS and evaluate candidate biomarkers in public datasets and small independent human samples, with supportive mouse expression evidence.
Baseline multisystem biomarkers were associated with subsequent cardiovascular events in an observational analysis. Adjusted associations and statistical mediation do not establish prevention by lowering the score or an inflammatory marker, or a validated clinical prediction model.
Plasma Olink proteomics for biomarker discovery and diagnosis in dilated cardiomyopathy with heart failure.
A two-center exploratory study profiled 92 circulating EV-associated proteins in AAA and evaluated five candidates by Western blot and ELISA. External sample size, assay-dependent expression differences and wide ROC uncertainty are retained explicitly.
Transcriptomic CRDscore and coronary-disease subtype analysis, with explicit qualification of repeated samples and reference-dataset identity.
A secondary linked-mortality analysis relates self-reported sleep to mortality in a complex-survey population with hyperuricemia. Associations and the fitted turning point are epidemiological estimates, not a survival-improving sleep prescription; sleep-trouble findings do not cover every diagnosed sleep disorder.
This methodological Perspective argues for context-sensitive learner integration, evaluation across independent cohorts and cautious interpretation of feature attribution. Its deployment discussion identifies requirements for future clinical translation rather than reporting a newly validated model.
This Correspondence discusses a research agenda connecting multimodal biological interpretation with patient-specific simulation in cardiovascular medicine. It sets out design and translation requirements without reporting a newly built, clinically validated digital twin system.
A mechanistic study linking idebenone-mediated cardioprotection after myocardial infarction to ROS, autophagy, ferroptosis and AMPK–mTOR signaling.
Recording results from single-patient trials
Read the evidenceAI hybrid survival assessment for advanced heart failure patients with renal dysfunction
Read the evidenceSystem biology analysis reveals circadian rhythm disorder associated with development and progression in colorectal cancer
Read the evidenceGe Zhang is a clinician-scientist working at the intersection of AI for Science, systems medicine, and translational biomedicine. His research integrates explainable artificial intelligence, multimodal clinical data, and multi-omics profiling to resolve disease heterogeneity, uncover mechanistic determinants, and translate molecular signatures into clinically actionable knowledge across major chronic diseases. Cardiovascular disease, particularly atherosclerosis, represents a major disease context in his research, with specific interests in the vascular microenvironment, vascular smooth muscle cell plasticity, and circadian disruption. His work spans population-scale clinical modeling, multimodal and multi-omics integration, computational inference, mechanistic interrogation, and translational validation, establishing cross-scale links from molecular and cellular states to clinical phenotypes and disease outcomes.
张格（Ge Zhang），临床医学与计算生物医学研究者，主要从事 AI for Science、系统医学与转化生物医学研究。其研究面向重大慢性疾病，融合可解释人工智能、多模态临床数据与多组学研究体系，旨在解析疾病异质性、识别关键机制驱动因素，并推动分子特征向具有临床意义的生物学认知与转化标志物延伸。 心血管疾病，尤其是动脉粥样硬化，是其重要疾病研究场景，重点关注动脉粥样硬化血管微环境、血管平滑肌细胞可塑性及昼夜节律紊乱。 其研究贯穿人群尺度临床建模、多模态与多组学整合、计算推断、机制解析与转化验证，致力于建立从分子与细胞状态到临床表型及疾病结局的跨尺度关联。
Author-controlled pages connect publications, research concepts, persistent identifiers and code/data resources.
Open to research collaborations and academic exchange.
