InstaDeepAI/nucleotide_transformer_downstream_tasks_revised

核苷酸转换器下游任务数据集包含了18个下游任务，这些任务来自于核苷酸转换器论文中提出的基因组基准测试。任务包括二分类和多分类任务，旨在提供一致的基因组学基准。数据集更新后，所有数据均来自人类样本，并且是来自近期的高质量数据集。数据集中的负样本由现有的不包含相应元素的基因组片段替换，且所有数据集现在都有适当的染色体保留测试集。

The Nucleotide Transformer downstream tasks dataset includes 18 downstream tasks presented in the Nucleotide Transformer paper, which consist of binary and multi-class classification tasks aiming to provide a consistent genomics benchmark. After the update, all data is from human samples and from recent high-quality datasets. Negative samples are replaced by existing genome chunks not containing the respective elements, and all datasets now have proper chromosome held-out test sets.