Only 6% of known species have a conservation status. Methods that assess conservation statuses are often based on individual counts and are thus too laborious to be generalized to all species. Population genomics methods that infer past variations in population size are easy to use but limited to the relatively distant past. Here we propose a population genomics approach that tests for recent population decline and may be used to assess species conservation statuses. More specifically, we study Maximal Recombination Free (MRF) blocks, that are segments of a sequence alignment inherited from a common ancestor without recombination. MRF blocks are relatively longer in small than in large populations. We use the distribution of MRF block lengths rescaled by their mean to test for recent population decline. However, because MRF blocks are difficult to detect, we also consider Maximal Linkage Disequilibrium (MLD) blocks, which are runs of single nucleotide polymorphisms compatible with a single tree. We develop a new method capable of inferring a very recent decline (e.g. with a detection power of 50% for populations whose size was halved to $N$, 0.05 $\times N$ generations ago) from rescaled MLD block lengths. Our framework could serve as a basis for quantitative tools to assess conservation status in a wide range of species.

只有6％的已知物种具有保护状态。评估保护状态的方法通常基于个体计数，因此过于费力，无法推广到所有物种。推断过去种群规模变化的种群基因组学方法易于使用，但仅限于相对较远的过去。在这里，我们提出了一种人口基因组学方法，该方法可以测试最近的种群下降情况，并可以用来评估物种的保护状况。更具体地说，我们研究了最大无重组（MRF）块，它们是从一个共同祖先继承而无需重组的序列比对的片段。小型地区的MRF区块比大型地区的MRF区块相对较长。我们使用按均值重新缩放的MRF块长度分布来测试最近的人口下降。然而，因为MRF块很难检测，所以我们还考虑了最大连锁不平衡（MLD）块，它是与单个树兼容的单核苷酸多态性序列。我们开发了一种新方法，能够推断最近的下降趋势（例如，对于大小减半到$ñ$，0.05 $\times ñ$几代人之前的版本）。我们的框架可以作为定量工具的基础，以评估各种物种的保护状况。