bolinas-dna/zoonomia-v1-v1
收藏资源简介：
zoonomia-v1-v1是一个基因组学数据集，包含255 bp的人类锚定窗口，经过保守性过滤（使用phyloP_447m评分，保守比例≥0.20），并通过halLiftover投影到108个去家族重复的Zoonomia 447哺乳动物组装上，进行中点调整至255 bp，同时进行了反向互补增强。数据集仅包含一个训练分割，总行数为223,880,280。数据模式包括查询名称、物种、目标染色体、起始位置、结束位置、链方向、目标染色体大小、序列（固定255 bp，链感知）和增强类型（原始或反向互补）。该数据集适用于DNA序列分析和机器学习模型训练。
zoonomia-v1-v1 is a genomics dataset comprising 255 bp human-anchored windows that have been conservation-filtered (using phyloP_447m score with proportion_conserved >= 0.20) and projected onto 108 family-deduped Zoonomia 447-mammalian assemblies via halLiftover, with midpoint resizing to 255 bp and reverse-complement augmentation. It includes a single train split with 223,880,280 rows. The schema features columns such as query_name, species, t_chrom, t_start, t_end, t_strand, t_src_size, sequence (exactly 255 bp, strand-aware), and augmentation (original or reverse-complemented). This dataset is designed for DNA sequence analysis and machine learning model training.
