Cornman Robert S, Fike Jennifer A, Oyler-McCance Sara J, Cryan Paul M
U.S. Geological Survey, Fort Collins Science Center, Fort Collins, CO, United States of America.
PeerJ. 2021 Apr 19;9:e11285. doi: 10.7717/peerj.11285. eCollection 2021.
Hoary bats () are among the bat species most commonly killed by wind turbine strikes in the midwestern United States. The impact of this mortality on species census size is not understood, due in part to the difficulty of estimating population size for this highly migratory and elusive species. Genetic effective population size (Ne) could provide an index of changing census population size if other factors affecting Ne are stable.
We used the NeEstimator package to derive effective breeding population size (Nb) estimates for two temporally spaced cohorts: 93 hoary bats collected in 2009-2010 and an additional 93 collected in 2017-2018. We sequenced restriction-site associated polymorphisms and generated a de novo genome assembly to guide the removal of sex-linked and multi-copy loci, as well as identify physically linked markers.
Analysis of the reference genome with suggested at least a doubling of Ne in the last 100,000 years, likely exceeding Ne = 10,000 in the Holocene. Allele and genotype frequency analyses confirmed that the two cohorts were comparable, although some samples had unusually high or low observed heterozygosities. Additionally, the older cohort had lower mean coverage and greater variability in coverage, and batch effects of sampling locality were observed that were consistent with sample degradation. We therefore excluded samples with low coverage or outlier heterozygosity, as well as loci with sequence coverage far from the mode value, from the final data set. Prior to excluding these outliers, contemporary Nb estimates were significantly higher in the more recent cohort, but this finding was driven by high values for the 2018 sample year and low values for all other years. In the reduced data set, Nb did not differ significantly between cohorts. We found base substitutions to be strongly biased toward cytosine to thymine or the complement, and further partitioning loci by substitution type had a strong effect on Nb estimates. Minor allele frequency and base quality bias thresholds also had strong effects on Nb estimates. Instability of Nb with respect to common data filtering parameters and empirically identified factors prevented robust comparison of the two cohorts. Given that confidence intervals frequently included infinity as the stringency of data filtering increased, contemporary trends in Nb of North American hoary bats may not be tractable with the linkage disequilibrium method, at least using the protocol employed here.
霜蝠（）是美国中西部地区因风力涡轮机撞击而最常死亡的蝙蝠物种之一。这种死亡率对物种普查规模的影响尚不清楚，部分原因是难以估计这种高度迁徙且难以捉摸的物种的种群规模。如果影响有效种群大小（Ne）的其他因素稳定，遗传有效种群大小（Ne）可以提供种群普查规模变化的指标。
我们使用NeEstimator软件包来推导两个时间间隔的队列的有效繁殖种群大小（Nb）估计值：2009 - 2010年收集的93只霜蝠以及2017 - 2018年收集的另外93只。我们对限制性位点相关多态性进行测序，并生成了一个从头基因组组装，以指导去除性连锁和多拷贝位点，以及识别物理连锁的标记。
对参考基因组的分析表明，在过去10万年中Ne至少增加了一倍，在全新世可能超过Ne = 10,000。等位基因和基因型频率分析证实两个队列具有可比性，尽管一些样本观察到的杂合度异常高或低。此外，较老的队列平均覆盖率较低且覆盖率变异性较大，并且观察到采样地点的批次效应与样本降解一致。因此，我们从最终数据集中排除了覆盖率低或杂合度异常的样本，以及序列覆盖率远离模式值的位点。在排除这些异常值之前，当代Nb估计值在较新的队列中显著更高，但这一发现是由2018年样本年的高值和所有其他年份的低值驱动的。在简化数据集中，两个队列之间的Nb没有显著差异。我们发现碱基替换强烈偏向于胞嘧啶到胸腺嘧啶或其互补方向，并且按替换类型进一步划分位点对Nb估计值有很大影响。次要等位基因频率和碱基质量偏差阈值对Nb估计值也有很大影响。Nb相对于常见数据过滤参数和经验确定因素的不稳定性使得无法对两个队列进行可靠比较。鉴于随着数据过滤的严格性增加，置信区间经常包含无穷大，至少使用此处采用的方案，北美霜蝠当代Nb的趋势可能无法通过连锁不平衡方法来处理。
