Biomarker & Target discovery
Biomarker & Target discovery
Discover disease-driving targets and responder biomarkers directly from patient data.
ConvergeCELL™ is our virtual cell and precision medicine solution. It supports target discovery and biomarker discovery by learning from cell and patient-level gene expression data. It detects and compares expression patterns and cell-type-specific signatures across patient populations to highlight genes linked to treatment response.
ConvergeCELL™ is our virtual cell and precision medicine solution. It supports target discovery and biomarker discovery by learning from cell and patient-level gene expression data. It detects and compares expression patterns and cell-type-specific signatures across patient populations to highlight genes linked to treatment response.
01
01
Knowledge transfer from large biology
Pretrained on diverse single-cell atlases, ConvergeCELL stabilizes and denoises new, limited datasets.
Pretrained on diverse single-cell atlases, ConvergeCELL stabilizes and denoises new, limited datasets.
02
02
Richer bulk representations
ConvergeCELL encodes bulk samples using representations learned from single-cell data, capturing higher-order biological structure.
ConvergeCELL encodes bulk samples using representations learned from single-cell data, capturing higher-order biological structure.
03
03
Signal vs. noise disentanglement
ConvergeCELL separates patient noise from coherent mechanistic subgroups with distinct biological signatures.
ConvergeCELL separates patient noise from coherent mechanistic subgroups with distinct biological signatures.
04
04
Mechanistically grounded outputs
ConvergeCELL links discoveries to curated pathways, literature, and biological evidence.
ConvergeCELL links discoveries to curated pathways, literature, and biological evidence.
How it works
How it works
ConvergeCELL is trained on a large, diverse multimodal dataset of over 23 million cells from 5,000 patient samples, 550 studies, and 40 clinical indications. This foundation enables robust target and biomarker discovery across disease areas, integrating single-cell and bulk transcriptomic/proteomic data to surface biologically and clinically meaningful insights.
Input
Input Data Integration
Customer expression data (Single-cell or Bulk, Transcriptomics or Proteomics) is processed alongside Converge's proprietary Patient Matrix.
Input Data Integration
Customer expression data (Single-cell or Bulk, Transcriptomics or Proteomics) is processed alongside Converge's proprietary Patient Matrix.
Contrastive Learning
Supervised Contrastive Learning (CL)
A Cell Encoder and Cell Aggregator learn a high-fidelity, unified representation for each patient from their single-cell data, capturing the full biological context rather than disconnected cell signals.
Supervised Contrastive Learning (CL)
A Cell Encoder and Cell Aggregator learn a high-fidelity, unified representation for each patient from their single-cell data, capturing the full biological context rather than disconnected cell signals.
Explainability
Explainability & Output
A Classification Head links the patient representation to clinical outcomes. The model then extracts Disease-specific association scores and Cell-type-gene importance scores. This explainability allows us to extract the genes that are the main drivers for a specific disease or response.
Explainability & Output
A Classification Head links the patient representation to clinical outcomes. The model then extracts Disease-specific association scores and Cell-type-gene importance scores. This explainability allows us to extract the genes that are the main drivers for a specific disease or response.
Output
Actionable Insights
This process yields a prioritized list of known targets (for validation) and Novel Targets (for first-in-class programs), each with a traceable, evidence-based rationale.
Actionable Insights
This process yields a prioritized list of known targets (for validation) and Novel Targets (for first-in-class programs), each with a traceable, evidence-based rationale.
Frequently asked questions
Frequently asked questions
How is ConvergeCELL™ validated and benchmarked?
What is the minimum amount of data required?
How is ConvergeCELL™ different from traditional differential expression or ML approaches?
What inputs are required to get started with ConvergeCELL™?
What outputs does ConvergeCELL™ provide?
How is ConvergeCELL™ validated and benchmarked?
What is the minimum amount of data required?
How is ConvergeCELL™ different from traditional differential expression or ML approaches?
What inputs are required to get started with ConvergeCELL™?
What outputs does ConvergeCELL™ provide?
Accelerate your
target and biomarker discovery.
Accelerate your
target and biomarker discovery.