scpFormer: A Foundation Model for Unified Representation and Integration of the Single-Cell Proteomics

2026-04-21

AI Summary

Single-cell proteomic data are challenging to integrate due to the fragmented nature of targeted antibody panels. To address this, this work proposes scpFormer, a Transformer-based foundation model that maps variable antibody panels into a unified semantic space through continuous sequence anchoring embeddings, evolutionary-scale modeling, and expression-aware encoding. Leveraging panel-agnostic continuous tokenization and an open-vocabulary architecture, scpFormer enables unsupervised clustering, cross-batch integration, and in silico panel expansion, while effectively reconstructing biological manifolds even from sparse clinical samples. Experimental results demonstrate that scpFormer excels in large-scale integration tasks and successfully facilitates precision oncology applications, including cancer drug response prediction.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems
• Computer Vision: Medical and Biological Imaging

Application Category

• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Security and Privacy: Data transparency and provenance
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

The integration of single-cell proteomic data is often hindered by the fragmented nature of targeted antibody panels. To address this limitation, we introduce scpFormer, a transformer-based foundation model designed for single-cell proteomics. Pre-trained on over 390 million cells, scpFormer replaces standard index-based tokenization with a continuous, sequence-anchored approach. By combining Evolutionary Scale Modeling (ESM) with value-aware expression embeddings, it dynamically maps variable panels into a shared semantic space without artificial discretization. We demonstrate that scpFormer generates global cell representations that perform competitively in large-scale batch integration and unsupervised clustering. Moreover, its open-vocabulary architecture facilitates in silico panel expansion, assisting in the reconstruction of biological manifolds in sparse clinical datasets. Finally, this learned protein co-expression logic is transferable to bulk-omics tasks, supporting applications like cancer drug response prediction. scpFormer provides a versatile, panel-agnostic framework to facilitate scalable biomarker discovery and precision oncology.

Problem

• single-cell proteomics
• data integration
• antibody panels
• fragmentation
• unified representation

Innovation

• foundation model
• single-cell proteomics
• continuous tokenization
• panel-agnostic integration
• protein co-expression