Genome Understanding Evaluation (GUE)
收藏资源简介:
Genome Understanding Evaluation (GUE)数据集是由西北大学计算机科学系的研究团队开发的,旨在为基因组理解提供一个标准化的评估平台。该数据集包含36个子数据集,覆盖了从人类、真菌、病毒到酵母等多种物种的基因组分析任务,输入序列长度从70到10000不等。GUE数据集通过精心设计的策略确保了数据集的难度适中,能够准确反映现有基因组基础模型的能力。该数据集的应用领域广泛,旨在解决基因组序列分析中的多种问题,如转录因子预测、增强子-启动子相互作用预测等,为基因组学研究提供了强大的工具和基准。
The Genome Understanding Evaluation (GUE) dataset was developed by a research team from the Department of Computer Science at Northwestern University, aiming to provide a standardized assessment platform for genome understanding. This dataset includes 36 sub-datasets, covering genome analysis tasks across diverse species such as humans, fungi, viruses, yeast and more. The lengths of the input sequences range from 70 to 10,000. The GUE dataset adopts carefully designed strategies to ensure a moderate level of difficulty, enabling it to accurately reflect the capabilities of current genomic foundation models. It has a wide range of application scenarios, targeting multiple problems in genomic sequence analysis including transcription factor prediction and enhancer-promoter interaction prediction, serving as a powerful tool and benchmark for genomics research.




