One important task in single-cell analysis is to quantify the differentiation potential of single cells. Though various single-cell potency measures have been proposed, they are based on individual biological sources, thus not robust and reliable. It is still a challenge to combine multiple sources to generate a relatively reliable and robust measure to estimate differentiation. In this paper, we propose a New Centrality measure with Gene ontology information (NCG) to estimate single-cell potency. NCG is designed by combining network topology property with edge clustering coefficient, and gene function information using gene ontology function similarity scores. NCG distinguishes pluripotent cells from non-pluripotent cells with high accuracy, correctly ranks different cell types by their differentiation potency, tracks changes during the differentiation process, and constructs the lineage trajectory from human myoblasts into skeletal muscle cells. These indicate that NCG is a reliable and robust measure to estimate single-cell potency. NCG is anticipated to be a useful tool for identifying novel stem or progenitor cell phenotypes from single-cell RNA-Seq data. The source codes and datasets are available at https://github.com/Xinzhe-Ni/NCG .

中文翻译：

---

基于网络拓扑和基因本体信息准确估计单细胞分化潜能

单细胞分析的一项重要任务是量化单细胞的分化潜能。尽管已经提出了各种单细胞效能测量方法，但它们都是基于个体生物来源，因此不够稳健和可靠。结合多个来源以生成相对可靠和稳健的方法来估计差异化仍然是一个挑战。在本文中，我们提出了一种使用基因本体信息 (NCG) 的新中心性度量来估计单细胞效能。NCG是通过将网络拓扑特性与边缘聚类系数相结合，以及使用基因本体功能相似性得分的基因功能信息来设计的。NCG 可高精度区分多能细胞和非多能细胞，根据分化能力对不同细胞类型进行正确排序，跟踪分化过程中的变化，并构建从人类成肌细胞到骨骼肌细胞的谱系轨迹。这些表明 NCG 是估计单细胞效能的可靠且稳健的措施。NCG 有望成为从单细胞 RNA-Seq 数据中识别新的干细胞或祖细胞表型的有用工具。源代码和数据集可在https://github.com/Xinzhe-Ni/NCG .