Cancer Research UK Cambridge Institute, University of Cambridge, Li Ka Shing Centre, Robinson Way, Cambridge, CB2 0RE, UK.
Genome Med. 2024 Nov 26;16(1):139. doi: 10.1186/s13073-024-01414-4.
CRISPR-Cas9 technology has revolutionised genetic screens and can inform on gene essentiality and chemo-genetic interactions. It is easily deployed and widely supported with many pooled CRISPR libraries available commercially. However, discrepancies between the reference genomes used in the design of those CRISPR libraries and the cell line under investigation can lead to loss of signal or introduction of bias. The problem is particularly acute when dealing with variant cell lines such as cancer cell lines.
Here, we present an algorithm, EXOme-guided Re-annotation of nuCleotIde SEquences (Exorcise), which uses sequence search to detect and correct mis-annotations in CRISPR libraries. Exorcise verifies the presence of CRISPR targets in the target genome and applies corrections to CRISPR libraries using existing exome annotations. We applied Exorcise to re-annotate guides in pooled CRISPR libraries available on Addgene and found that libraries designed on a more permissive reference sequence had more mis-annotations. In simulated CRISPR screens, we modelled common mis-annotations and found that they adversely affect discovery of hits in the intermediate range. We then confirmed this by applying Exorcise on datasets from Dependency Map (DepMap) and the DNA Damage Response CRISPR Screen Viewer (DDRcs), where we found improved discovery power upon Exorcise while retaining the strongest hits.
Pooled CRISPR libraries map guide sequences to genes and these mappings might not be ready to use due to permissive library design or investigating a variant cell line. By re-annotating CRISPR guides, Exorcise focuses CRISPR experiments towards the genome of the cell line under investigation. Exorcise can be applied at the library design stage or the analysis stage and allows post hoc re-analysis of completed screens. It is available under a Creative Commons Zero v1.0 Universal licence at https://github.com/SimonLammmm/exorcise .
CRISPR-Cas9 技术彻底改变了遗传筛选,可以提供有关基因必需性和化学生物学相互作用的信息。它易于部署,并且得到了广泛支持,许多商业上可用的 pooled CRISPR 文库。然而,这些 CRISPR 文库设计中使用的参考基因组与被研究的细胞系之间的差异可能导致信号丢失或引入偏差。当处理变体细胞系(如癌细胞系)时,问题尤其严重。
在这里,我们提出了一种算法,即 EXOme-guided Re-annotation of nuCleotIde SEquences(Exorcise),它使用序列搜索来检测和纠正 CRISPR 文库中的错误注释。Exorcise 验证了靶基因组中 CRISPR 靶标的存在,并使用现有的外显子注释对 CRISPR 文库进行校正。我们应用 Exorcise 重新注释 Addgene 上可用的 pooled CRISPR 文库中的指南,发现设计时使用更宽松的参考序列的文库有更多的错误注释。在模拟的 CRISPR 筛选中,我们模拟了常见的错误注释,发现它们会对中等范围内的命中发现产生不利影响。然后,我们通过在 Dependency Map(DepMap)和 DNA 损伤反应 CRISPR 筛选查看器(DDRcs)上应用 Exorcise 来证实这一点,在这些数据集上应用 Exorcise 后,发现发现能力得到了提高,同时保留了最强的命中。
pooled CRISPR 文库将引导序列映射到基因上,并且由于库设计的宽容性或研究变体细胞系,这些映射可能无法直接使用。通过重新注释 CRISPR 指南,Exorcise 将 CRISPR 实验集中在被研究的细胞系的基因组上。Exorcise 可以在库设计阶段或分析阶段应用,并允许对已完成的筛选进行事后重新分析。它在 Creative Commons Zero v1.0 Universal 许可证下可在 https://github.com/SimonLammmm/exorcise 获得。