arXiv:2510.24987v1 Announce Type: cross 
Abstract: Advances in single-cell sequencing have enabled high-resolution profiling of diverse molecular modalities, while integrating unpaired multi-omics single-cell data remains challenging. Existing approaches either rely on pair information or prior correspondences, or require computing a global pairwise coupling matrix, limiting their scalability and flexibility. In this paper, we introduce a scalable and flexible generative framework called single-cell Multi-omics Regularized Disentangled Representations (scMRDR) for unpaired multi-omics integration. Specifically, we disentangle each cell's latent representations into modality-shared and modality-specific components using a well-designed $\beta$-VAE architecture, which are augmented with isometric regularization to preserve intra-omics biological heterogeneity, adversarial objective to encourage cross-modal alignment, and masked reconstruction loss strategy to address the issue of missing features across modalities. Our method achieves excellent performance on benchmark datasets in terms of batch correction, modality alignment, and biological signal preservation. Crucially, it scales effectively to large-level datasets and supports integration of more than two omics, offering a powerful and flexible solution for large-scale multi-omics data integration and downstream biological discovery.

تقديم scMRDR يمثل تقدمًا كبيرًا في دمج بيانات المولتي أومكس أحادية الخلية غير المتزاوجة. هذا الإطار الجديد يعالج قيود الأساليب الحالية، التي تعتمد غالبًا على المعلومات المتزاوجة أو الحسابات المعقدة، مما يحد من قابليتها للتوسع. من خلال تقديم نهج أكثر مرونة، يفتح scMRDR آفاقًا جديدة للباحثين في هذا المجال، مما يسمح بإجراء تحليلات أكثر شمولاً للملفات الجزيئية. هذه الابتكار مهم لأنه يعزز قدرتنا على فهم الأنظمة البيولوجية المعقدة على مستوى الخلية الواحدة.

La introducción de scMRDR marca un avance significativo en la integración de datos multi-ómicos unicelulares no emparejados. Este nuevo marco aborda las limitaciones de los métodos existentes, que a menudo dependen de información emparejada o cálculos complejos, lo que limita su escalabilidad. Al ofrecer un enfoque más flexible, scMRDR abre nuevas posibilidades para los investigadores en el campo, permitiendo análisis más completos de los perfiles moleculares. Esta innovación es crucial, ya que mejora nuestra capacidad para entender sistemas biológicos complejos a nivel unicelular.

L'introduction de scMRDR représente une avancée significative dans l'intégration des données multi-omiques unicellulaires non appariées. Ce nouveau cadre répond aux limitations des méthodes existantes, qui dépendent souvent d'informations appariées ou de calculs complexes, ce qui les rend moins évolutives. En offrant une approche plus flexible, scMRDR ouvre de nouvelles possibilités pour les chercheurs dans le domaine, permettant des analyses plus complètes des profils moléculaires. Cette innovation est cruciale car elle améliore notre capacité à comprendre des systèmes biologiques complexes au niveau unicellulaire.

The introduction of scMRDR marks a significant advancement in the integration of unpaired single-cell multi-omics data. This new framework addresses the limitations of existing methods, which often depend on paired information or complex computations, making it less scalable. By offering a more flexible approach, scMRDR opens up new possibilities for researchers in the field, allowing for more comprehensive analyses of molecular profiles. This innovation is crucial as it enhances our ability to understand complex biological systems at a single-cell level.

scMRDR: A scalable and flexible framework for unpaired single-cell multi-omics data integration

One More Thing in AI – Your Shortcut to AI Mastery

scMRDR: A scalable and flexible framework for unpaired single-cell multi-omics data integration

Was this article worth reading? Share it

One More Thing in AI

Magicley AI

LucidQuery AI

Dynamiq

Metaflow AI

Supametas.AI

Ready to build your own newsroom?