表层 · 技术 · 技术

Immediately Available Dry Computing Services

Not PPT visions — but engineered capabilities that are tested, verified, and deliverable

Flagship

Neoantigen mRNA Vaccine 8-Step Pipeline

Verified

From VCF/BAM to mRNA sequence design, fully validated end-to-end. Three differentiators: pMHC 3D structure verification, 5-dimensional immunogenicity scoring, in-house RNALens fine-tuning. TESLA benchmark AUROC=0.698.

45 candidates · 17 Tier-1 · IC50 min 10.1 nM · pLDDT 95.4· Sample available
Core Differentiator

pMHC 3D Structure Prediction + Docking Validation

Verified

Traditional pipelines stop at IC50 values. DiVo Gen²AI uses Protenix to build atom-level pMHC 3D structures for each candidate neoantigen, validated by Protenix-Dock physical docking. 400+ complexes at scale.

pLDDT 95.4 · ipTM 0.977 · 5-group docking bscore -69~-92· Sample available
General Service

Genome Interpretation Report (Bilingual)

Verified

VCF QC → VEP annotation → ANNOVAR functional annotation → ACMG classification → health report. Bilingual MD/HTML/PDF delivery, 12+ batches validated.

12+ batches · Bilingual · MD/HTML/PDF· Sample available
Entry Service

HLA Typing (Dual-Tool Cross-Validation)

Verified

OptiType + Polysolver dual-algorithm cross-validation, 4-digit resolution HLA-I typing. Provides基础 input for neoantigen pipelines, also independently deliverable.

65+25 alleles · Dual-tool cross-validation· Sample available
In-House Fine-Tuned

mRNA Sequence Optimization (RNALens)

Verified

MRL prediction model fine-tuned in 3 stages on DNABERT-Z 117M — not a third-party API call. CAI 0.7→0.95, 3 tissue-adapted models (HEK/Muscle/PC3).

Spearman 0.92 · R² 0.87 · CAI 0.7→0.95· Sample available
New Service

CAR-T Cell Therapy Computation

Verified

Neoantigen pipeline + RNA-seq surface antigen mining + de novo Binder design + structure validation. For medical institution partnerships, not direct-to-patient. Xie Qi / Cao Longxing Nature BME路线.

Dual-channel targets · RFdiffusion+MPNN · pLDDT 95.4
Beauty & Anti-Aging

Aesthetic PRS · Genetic Beauty Profile

Verified

Polygenic risk scores for aesthetic targets — collagen loss, photoaging sensitivity, hair loss risk, pigmentation. PRSice-2+PRS-CS dual engine, Chinese population calibrated. B2B2C model — the tech backbone for aesthetic clinics.

5-target PRS · PRSice-2+PRS-CS · GVM+NyuWa calibrated
Beauty & Anti-Aging

Anti-Aging Peptide & Protein Engineering

Verified

Finding anti-aging molecules in 3D space — MMPs inhibitors, collagen engineering, Sirtuin activators. 100% tools deployed and validated: Protenix V2 + Pro-Prime + BoltzGen(5-weight complete) + ProteinMPNN + MHCflurry.

pLDDT 95.4 · Pro-Prime ΔΔG · BoltzGen 5-weight · 87 tools
Beauty & Anti-Aging

Skin Microbiome × Host Gene Joint Mining

Verified

Skin condition is a symbiotic result of genes and microbiota. 16S/ITS microbiome analysis (DADA2 local + MGnify online fallback) cross-mined with host HLA/FLG immune genes, revealing gene-microbiota-skin phenotype ternary relationships.

DADA2+MGnify · ANNOVAR+HLA validated · SILVA 138.2
Beauty & Anti-Aging

Follicle Induction · Five-Factor Combination

Exploratory

From repair to regeneration — WNT10B+FGF9+Noggin+RSPO1+SHH-N five-factor combination inducing new follicles on hairless skin. Protenix V2 structure prediction + ProteinMPNN engineering design validated; mRNA delivery design (GEMORNA) ready. ARM mode: design only, no manufacturing.

Protenix pLDDT 95.4 · 5-factor structure predictable · mRNA delivery ready
Verified9
Exploratory1
Conceptual0
View all 10 services
表层 · 技术 · 技术

ARCHITECTURE

四层业务能力架构

从硅片到信息载体 · 四层堆叠 · 四类用户各取所需

L1-L3 的底层实现细节不在对外页面披露范围,可在合作框架下按需对接。

中层 · 方法 · 方法

METHOD

DiVo Computing · 砥尚计算

Distributed Virtuous Computing — 分布式、有优雅而坚固的计算体系

Agentic Computing 的螺旋

信息数据化计算数据再信息化 →↑ 理解的持续深化

不是单次计算,而是螺旋上升——信息数据化后进入计算,计算产出的数据再交付给人类时就是信息的载体,信息驱动更深的理解,更深的理解触发更精准的计算。DiVo Computing 是这个螺旋的工程化载体。

分布式 · Distributed

对抗结构刚性。单体算力有限、单点逻辑有偏——分布式体系让不同模型、不同工具、不同视角交叉验证,可靠性数量级提升。

有德性 · Virtuous

约束工具向善。计算不是目的,理解才是。我们拒绝制造 F_damage——不让技术成为伤害的放大器,让计算回归提升理解这一根本目的。

中层 · 方法 · 方法

开始你的理解之旅

从基因到蛋白质,从突变到疫苗,从数据到决策——DiVo Gen²AI 已经在跑。

DiVo Gen²AI · 前沿 + 全能 · DiVo Computing 砥尚计算 · Understanding is All You Need

底层 · 思想 · 思想

FOUNDATION

Understanding is All You Need.

帮助人类以科学的方式理解世界与自身,是我们全部的使命。

理解世界

从基因组到蛋白质结构,从肿瘤突变到药物设计——用计算让复杂变得可理解。

理解自身

每个人都是独特的——基因、免疫、代谢。真正的理解必须个体化,而非模板化。

理解即赋能

理解不是为了控制,是为了选择。让个体在科学理解的基础上做出更好的生命决策。