kuleshov-group/cross-species-single-nucleotide-annotation

该数据集用于跨物种植物基因组的单核苷酸分辨率建模，包含五个任务：翻译起始位点预测、翻译终止位点预测、剪接供体位点预测、剪接受体位点预测和进化保守性预测。训练数据集主要来自拟南芥和高粱的染色体，验证和测试数据集则分别来自拟南芥、水稻、高粱和玉米的染色体。数据集的大小和正负样本数量在README中有详细列出。

The dataset consists of five tasks for cross-species modeling plant genomes at single-nucleotide resolution in plants. These tasks include: 1. Translation Initiation Site (TIS) Prediction; 2. Translation Termination Site (TTS) Prediction; 3. Splice Donor Site Prediction; 4. Splice Acceptor Site Prediction; 5. Evolutionary Conservation Prediction. The training datasets for the first four tasks are generated from Arabidopsis chromosomes 1-4, the validation datasets from Arabidopsis chromosome 5, and the testing datasets compiled from rice, sorghum, and maize. The training dataset for the fifth task is generated from sorghum chromosomes 1-9, the validation dataset from sorghum chromosome 10, and the testing dataset compiled in maize. The dataset sizes are detailed with the number of positive and negative samples for each task.