depinwang/rnaseq-aligner-toy-benchmark-metrics-v1
收藏资源简介:
该数据集用于评估不同RNA-seq比对工具(star、hisat2、subjunc、bwa-mem2、minimap2)在模拟GRCh38 RNA-seq读取上的性能,特别是针对GENCODE v45转录本的真实内含子的剪接连接F1分数、精确度和召回率。数据集包含24行,每行代表一个比对工具在特定样本上的性能指标,包括真阳性、假阳性、假阴性、总调用连接数、总真实连接数等。数据集的来源包括实验rnaseq-aligner-toy-benchmark、集群Puhti (CSC)、参考基因组GRCh38.primary_assembly和GENCODE v45、模拟方法simulate_reads.py(均匀采样转录本,无错误模型)以及评分方法score_junctions.py(精确元组匹配,min_reads=3)。
This dataset evaluates the performance of different RNA-seq aligners (star, hisat2, subjunc, bwa-mem2, minimap2) on simulated GRCh38 RNA-seq reads, specifically focusing on splice-junction F1 / precision / recall against ground-truth introns from GENCODE v45 transcripts. The dataset contains 24 rows, each representing the performance metrics of an aligner on a specific sample, including true positives, false positives, false negatives, total junctions called, total junctions in ground-truth set, etc. The datasets provenance includes the experiment rnaseq-aligner-toy-benchmark, cluster Puhti (CSC), reference GRCh38.primary_assembly and GENCODE v45, simulation method simulate_reads.py (transcripts uniform-sampled, no error model), and scoring method score_junctions.py (exact tuple match, min_reads=3).




