scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing

Xu, Ping; Li, Pengjiang; Du, Tian; Wang, Zaitian; Gu, Jiawei; Ning, Zhiyuan; Qiao, Ziyue; Wang, Pengfei; Zhou, Yuanchun

doi:10.24963/ijcai.2026/350

scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing

Ping Xu, Pengjiang Li, Tian Du, Zaitian Wang, Jiawei Gu, Zhiyuan Ning, Ziyue Qiao, Pengfei Wang, Yuanchun Zhou

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Main Track. Pages 3151-3159.

https://doi.org/10.24963/ijcai.2026/350

Keywords

Data Mining: Applications
Machine Learning: Applications
Multidisciplinary Topics and Applications: Bioinformatics

Abstract

Clustering is fundamental to scRNA-seq analysis, serving as a cornerstone for identifying cell populations and resolving tissue heterogeneity. However, existing methods focus on mining numerical statistical patterns, suffering from semantic agnosticism by neglecting the intrinsic biological functions encoded by genes. While Large Language Models (LLMs) offer promising semantic capabilities, their direct adaptation to cell clustering is hindered by the structural mismatch between generative pre-training objectives and discriminative downstream tasks. To bridge this gap, we propose scLLM-DSC, a novel LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering framework. Diverging from data-driven paradigms, scLLM-DSC establishes a semantically-grounded representation by synergizing two views: a Knowledge-Driven Semantic View derived from NCBI gene priors and contextualized Cell2Sentence embeddings, and a Structure-Aware Topological View extracted via a graph-guided encoder. Crucially, we introduce a cross-modal contrastive alignment mechanism to enforce consistency between biological semantics and transcriptomic features within a unified latent space. Extensive benchmarks demonstrate that scLLM-DSC significantly outperforms eleven state-of-the-art baselines in clustering accuracy.

Ping Xu, Pengjiang Li, Tian Du, Zaitian Wang, Jiawei Gu, Zhiyuan Ning, Ziyue Qiao, Pengfei Wang, and Yuanchun Zhou. 2026. scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 3151–3159. https://doi.org/10.24963/ijcai.2026/350

@inproceedings{ijcai2026p350,
  title     = {scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing},
  author    = {Xu, Ping and Li, Pengjiang and Du, Tian and Wang, Zaitian and Gu, Jiawei and Ning, Zhiyuan and Qiao, Ziyue and Wang, Pengfei and Zhou, Yuanchun},
  booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-26}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Diego Calvanese},
  pages     = {3151--3159},
  year      = {2026},
  month     = {8},
  note      = {Main Track},
  doi       = {10.24963/ijcai.2026/350},
  url       = {https://doi.org/10.24963/ijcai.2026/350},
}

TY  - CPAPER
TI  - scLLM-DSC: LLM-Knowledge Enhanced Cross-Modal Deep Structural Clustering for Single-Cell RNA Sequencing
AU  - Xu, Ping
AU  - Li, Pengjiang
AU  - Du, Tian
AU  - Wang, Zaitian
AU  - Gu, Jiawei
AU  - Ning, Zhiyuan
AU  - Qiao, Ziyue
AU  - Wang, Pengfei
AU  - Zhou, Yuanchun
T2  - Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2026
SP  - 3151
EP  - 3159
DO  - 10.24963/ijcai.2026/350
UR  - https://doi.org/10.24963/ijcai.2026/350
ER  -