Bulk genotyping of biopsies can create spurious evidence for heterogeneity in mutation content

When multiple samples are taken from the neoplastic tissues of a single patient, it is natural to compare their mutation content. This is often done by bulk genotyping of whole biopsies, but the chance that a mutation will be detected in bulk genotyping depends on its local frequency in the sample. When the underlying mutation count per cell is equal, homogenous biopsies will have more high-frequency mutations, and thus more detectable mutations, than heterogeneous ones. Using simulations, we show that bulk genotyping of data simulated under a neutral model of somatic evolution generates strong spurious evidence for non-neutrality, because the pattern of tissue growth systematically generates differences in biopsy heterogeneity. Any experiment which compares mutation content across bulk-genotyped biopsies may therefore suggest mutation rate or selection intensity variation even when these forces are absent. We discuss computational and experimental approaches for resolving this problem....

当从单一患者的肿瘤组织中采集多个样本时，比较它们的突变含量是很自然的做法。这通常通过对整个活检样本进行批量基因分型（bulk genotyping）来实现，但某一突变在批量基因分型中被检测到的概率取决于其在样本中的局部频率。当每个细胞的潜在突变数量相同时，均质活检样本（homogenous biopsies）比异质活检样本（heterogeneous biopsies）具有更多的高频突变，因此可检测到的突变也更多。通过模拟，我们发现，对体细胞进化的中性模型（neutral model of somatic evolution）下模拟的数据进行批量基因分型会产生强烈的非中性（non-neutrality）虚假证据，因为组织生长模式会系统性地导致活检样本异质性的差异。因此，任何比较批量基因分型活检样本突变含量的实验，即使在不存在这些因素的情况下，也可能显示出突变率或选择强度变化（selection intensity variation）的迹象。我们讨论了解决这一问题的计算和实验方法……