Next generation biodiversity analysis

分类学是生物科学最古老的领域之一：识别和描述生物物种及其相互关系和环境。自从250年前林奈第一次开始这项任务以来，分类学家已经成功地描述了大约200万个地球物种。这些物种描述现在在各种各样的应用中发挥着关键作用，从环境生物监测到医学到农业。任何监测计划的一个基本原则是，人们无法控制无法测量的东西，对地球生物多样性总数的最佳估计从1000万到1亿多不等。显然，科学家们缺乏必要的工具来对生物多样性进行准确而有力的估计。此外，即使在存在正式物种描述的情况下，物种识别通常也需要大量的时间、精力和专业知识。基于DNA的方法有望改变这一切：通过分析短的、标准化的DNA片段——DNA条形码——有可能以极高的精度唯一地识别地球上几乎所有的物种。识别生活在特定生物群落中的每一个物种，并监测它们随时间的相对丰度，这在十年前是科幻小说里的东西，但通过将下一代DNA测序的能力与已经产生的庞大而多样的参考序列文库相结合，像这样雄心勃勃的研究现在是可能的——而且价格适中。我们建议利用计算机科学的进步，利用计算分析在标准文库中存储和分析这些DNA序列，然后用于监测任何环境下的物种多样性。我们从两个模型系统开始。第一种是淡水无脊椎动物群落——通常用于环境分析，如加拿大环境部进行的那些分析。我们的第二个模型系统涉及量化热带节肢动物的生物多样性，这一地区被广泛认为蕴藏着地球上大部分未知的生物多样性。

One of the oldest fields of biological science is taxonomy: the identification and description of biological species and their relationships to each other and their environments. Since Linnaeus first started the task 250 years ago, taxonomists have managed to describe roughly 2 million of the Earth's species. These species descriptions now play a key role in a wide variety of applications, from environmental biomonitoring to medicine to agriculture. A fundamental tenet of any monitoring program is that one cannot control what isn't measured, and the best estimates of the totality of Earth's biodiversity ranges from 10 million species to over 100 million. Clearly, scientists lack the required tools to create accurate and robust estimates of biodiversity. Moreover, species identification usually requires a lot of time, effort, and expertise even in those cases where formal species descriptions exist. DNA based approaches are poised to change all this: by analyzing short, standardized pieces of DNA-DNA barcodes--it is possible to uniquely identify nearly all the species on Earth with an extremely high degree of accuracy. Identifying every species living within a particular biome and monitoring their relative abundances over time would have been the stuff of science fiction ten years ago, but by combining the power of next-generation DNA sequencing with the large and diverse reference sequence libraries that have been produced, ambitious studies like these are now possible-and for a modest price. We propose to take advantage of advances in computer science and use computational analyses to store and analyze these DNA sequences in standard libraries that will then be used for monitoring species diversity in any environmental setting. We begin with two model systems. The first is freshwater invertebrate communities-the sort typically used in environmental analyses such as those performed by Environment Canada. Our second model system involves quantifying biodiversity among the tropical arthropods, an area that is widely believed to harbour the majority of the Earth's unknown biodiversity.