University of Veterinary Medicine Vienna, Vienna, Austria.
Vienna Graduate School of Population Genetics, Vienna, Austria.
BMC Bioinformatics. 2023 Aug 26;24(1):322. doi: 10.1186/s12859-023-05437-3.
The identification of genomic regions affected by selection is one of the most important goals in population genetics. If temporal data are available, allele frequency changes at SNP positions are often used for this purpose. Here we provide a new testing approach that uses haplotype frequencies instead of allele frequencies.
Using simulated data, we show that compared to SNP based test, our approach has higher power, especially when the number of candidate haplotypes is small or moderate. To improve power when the number of haplotypes is large, we investigate methods to combine them with a moderate number of haplotype subsets. Haplotype frequencies can often be recovered with less noise than SNP frequencies, especially under pool sequencing, giving our test an additional advantage. Furthermore, spurious outlier SNPs may lead to false positives, a problem usually not encountered when working with haplotypes. Post hoc tests for the number of selected haplotypes and for differences between their selection coefficients are also provided for a better understanding of the underlying selection dynamics. An application on a real data set further illustrates the performance benefits.
Due to less multiple testing correction and noise reduction, haplotype based testing is able to outperform SNP based tests in terms of power in most scenarios.
鉴定受选择影响的基因组区域是群体遗传学中最重要的目标之一。如果有时间序列数据，通常会使用 SNP 位置的等位基因频率变化来实现这一目标。本文提供了一种新的测试方法，该方法使用单倍型频率代替等位基因频率。
使用模拟数据，我们表明与基于 SNP 的测试相比，我们的方法具有更高的功效，尤其是当候选单倍型数量较少或中等时。为了在单倍型数量较大时提高功效，我们研究了将其与数量适中的单倍型子集相结合的方法。与 SNP 频率相比，单倍型频率通常可以以较少的噪声恢复，尤其是在池测序下，这使我们的测试具有额外的优势。此外，虚假的异常 SNP 可能导致假阳性，这在处理单倍型时通常不会遇到的问题。还提供了用于选择的单倍型数量和它们的选择系数之间差异的事后检验，以更好地了解潜在的选择动态。在真实数据集上的应用进一步说明了性能优势。
由于较少的多重检验校正和减少噪声，单倍型测试在大多数情况下在功效方面优于基于 SNP 的测试。
