计算机科学
人工智能
钥匙(锁)
特征(语言学)
光学(聚焦)
鉴定(生物学)
领域(数学)
组分(热力学)
背景(考古学)
机器学习
作者
Hong Xia,Boya Ji,Shaoliang Peng
出处
期刊:
[Institute of Electrical and Electronics Engineers]
日期:2026-04-21
卷期号:: 7452-7456
标识
DOI:10.1109/icassp55912.2026.11460643
摘要
The cell-cell communications (CCCs) have a significant role in the development of invasion and metastasis in tumor tissues. However, exploring CCCs using traditional experimental methods is time-consuming and costly. To assist with CCC analysis, a computational framework named CellExch is proposed. CellExch includes two main steps: construction of ligand-receptor interactions (LRIs) database and measurement of the strength of CCCs mediated by LRIs. Specifically, CellExch first predicts LRIs using multimodal features of ligands and receptors via graph convolutional networks and selects high-confident LRIs from these predictions. Then, CellExch quantifies the strength of cell–cell communication (CCC) by integrating the product result, thresholding result, and cell result obtained from high-confidence lig-and–receptor interactions (LRIs) and imputed gene expression data. A five-fold cross-validation was performed on each of the four benchmark datasets to evaluate the performance of CellExch, demonstrating its robustness in the LRI prediction task. Furthermore, we analyzed CCC in combination with high-confident LRIs in mouse kidney tissue. In summary, CellExch provides an effective LRI prediction model as well as an evaluation method for the strength of CCC, which allows researchers to perform CCC analysis in a more efficient manner.
科研通智能强力驱动
Strongly Powered by AbleSci AI