The big data revolution and human genetics.
The big data revolution and human genetics.
复制标题
大数据革命和人类遗传学。
DOI：
10.1093/hmg/ddy123
复制
发表时间：
2018
影响因子：
3.5
通讯作者：
Schork,NicholasJ
中科院分区：
文献类型：
--
作者：
Schork,NicholasJ
“大数据”的概念现在无处不在。几乎所有主要行业，无论是与金融、银行、营销、零售、社交媒体、能源或制造业相关的行业，都接受了对大数据集的分析，希望获得能够提高效率和创造更好产品的见解。因此，医疗保健、生物医学研究，特别是人类遗传学社区也接受大数据倡议也就不足为奇了。然而，在人类遗传学研究领域，创建和分析大规模数据集并不是什么新鲜事，因为一个人类基因组包含价值20亿美元的核苷酸信息。这些核苷酸是如何组织成基因及其相关的调控元件，导致特定功能，并相互作用，这是一项复杂而迷人的大数据分析工作，就像科学中的任何工作一样。然而，人类基因数据与其他数据类型的耦合或整合正变得越来越频繁，这实质上增加了人类基因研究人员愿意并渴望分析的已经非常庞大的数据集。本期《人类分子遗传学》致力于回顾挖掘人类基因组中固有的大数据，并将这些数据与其他数据类型相结合，以促进面向基因的生物医学科学和医疗保健的努力。鉴于DNA中的元素影响人类生理和调节致病过程的基本方式，正如这些综述所表明的那样，人类基因研究可以在无限多的环境中补充其他研究领域的数据，并与之结合。几乎所有的审查都考虑将基因数据与其他数据结合起来，以确定个人拥有的自然发生的遗传变异与各种表型之间的联系和联系，最明显的是那些可能具有临床和公共卫生效用的表型。Telenti和他的同事们考虑了生物信息学和数据分析工具的开发和应用，以解释人类基因组的变异，并表明通过结合数千个人类基因组进行分析，可以洞察遗传变异可能产生的功能影响。Scheuermann和他的同事们讨论了利用人类基因组中数十亿比特的信息来对人体内潜在的数万亿细胞进行编目、表征和细分的方法。范和他的同事们考虑将基因数据与成像数据，特别是神经成像数据结合起来，以获得仅靠遗传和成像数据无法获得的对人脑形态的洞察。考虑到当代人类遗传学研究的一个目标是将遗传信息纳入临床护理，将卫生系统中患者的基因数据与常规收集的临床信息结合起来是非常合乎逻辑的。不幸的是，将基因数据与常规临床数据结合起来充满了困难，因为临床数据通常是“噪音”的(例如，医生手写的笔记、输入患者记录的不同的临床参数测量方法等)。Altman和他的同事描述了确定基因变异和用于治疗各种疾病的药物的反应之间具有临床意义的关系的努力；而Wolford和他的同事、Diao和他的同事、Ohno-Machado和他的同事以及Glicksberg和他的同事都考虑对临床数据和基因数据进行一般综合分析，并着眼于根据从这些分析中获得的见解来改善健康。更进一步的是，Huentelman和Talom考虑将基因信息与数据整合到…
The concept of ‘Big Data’is now ubiquitous. Virtually all major industries, whether associated with finance, banking, marketing, retail, social media, energy or manufacturing, have embraced the analysis of big data sets hoping to obtain insights that could improve efficiency and create better products. It is no surprise then that the health care, biomedical research and, in particular, human genetics communities have also embraced big data initiatives. However, creating and analysing large-scale data sets is not particularly new in human genetics research contexts, since a single human genome contains 2Â $3.2 billion nucleotides worth of information. Just how these nucleotides are organized as genes and their associated regulatory elements, lead to particular functions, and interact, is as complicated and fascinating a big data analysis exercise as any in science. What is becoming more frequent, however, is the coupling or integration of human genetic data with other data types, essentially adding to the already very large data sets human genetic researchers are willing and eager to analyse. This issue of Human Molecular Genetics is devoted to reviews of efforts to both mine the big data inherent in human genomes and integrate that data with other data types to advance genetically oriented biomedical science and health care. Given the fundamental manner in which elements in DNA impact human physiology and mediate pathogenic processes, there are an unlimited number of settings in which human genetic research could complement, and be combined with, data from other research areas, as these reviews make clear. Virtually all of the reviews consider combining genetic data with other data to identify associations and connections between naturally occurring genetic variants possessed by individuals and phenotypes of all sorts, most notably those that may have clinical and public health utility. Telenti and colleagues consider the development and application of bioinformatics and data analysis tools to interpret variation in the human genome and show that by combining thousands of human genomes for analysis, insights into the likely functional effects of genetic variants can be found. Scheuermann and colleagues discuss methodology for leveraging what amounts to the billions of bits of information in the human genome to catalogue, characterize and subdivide the potentially trillions of cells in the human body. Fan and colleagues consider combining genetic data with imaging data, in particular neuroimaging data, in order to derive insights into human brain morphology that genetic and imaging data alone would not allow. Combining genetic data on patients in health systems with clinical information routinely collected on them is very logical given that a goal of contemporary human genetics research is to incorporate genetic information into clinical care. Unfortunately, combining genetic data with routine clinical data is fraught with difficulties given that clinical data are often ‘noisy’(eg physician hand-written notes, different ways of measuring clinical parameters entered into patient’s record, etc.). Altman and colleagues describe efforts to identify clinically meaningful relationships between genetic variants and responses to drugs used to treat various conditions; whereas Wolford and colleagues, Diao and colleagues, Ohno-Machado and colleagues and Glicksberg and colleagues all consider general integrated analysis of clinical data and genetic data with an eye towards improving health based on the insights obtained from such analyses. Pushing things even further, Huentelman and Talboom consider integrating genetic information with data …