What Can Networks Tell Us About Aging?

由于数以百万计的婴儿潮一代已经开始年满65岁，美国正在变老。由于疾病的易感性随着年龄的增长而增加，因此研究衰老的分子原因变得重要。人类的寿命很长，这除了伦理上的限制外，还使得研究人类衰老变得困难。因此，我们用更简单的模型来研究衰老。品种，如面包师？S酵母。然后，关于衰老的知识从模式物种转移到人类。到目前为止，这种转移仅限于基因组序列比较，通过识别不同物种基因序列之间的相似区域(被认为是序列之间功能关系的结果)，以及通过将知识从模式物种的基因转移到人类的序列相似基因。然而，基因(也就是它们的蛋白质产物)通过复杂的网络相互作用实现生物功能，而不是单独作用。因此，在后基因组时代，人们争辩说，细胞网络中基因之间的连接可以提供超越单个基因序列的生物学见解。因此，该项目假设，与基因组序列研究类似，生物网络研究将影响我们对衰老的理解。例如，由于在模式物种中并不是所有与衰老有关的基因在人类中都具有序列相似的基因，限制对序列的比较可能会限制与衰老相关的知识向人类的转移。网络比较可以有所帮助，因为它可以找到不同物种网络之间的相似区域，并允许这些区域之间的知识转移。智力优势：与基因组序列研究不同，生物网络研究处于起步阶段，原因如下。许多网络问题(包括网络比较)在计算上是棘手的，因此需要有效的近似(或启发式)解。许多基因的功能尚不清楚，因此，必须从其他特征更好的基因中发现它。尽管细胞随着时间的推移而进化，但目前用于分析系统级生物网络的方法只涉及它们的静态表示，这是因为利用现有的生物技术很难获得动态生物网络数据，并且因为缺乏有效的动态网络分析方法。由于生物技术的局限性以及人类在数据采集过程中的偏见，目前的生物网络存在噪声，有许多缺失和虚假的链接；因此，需要开发网络去噪方法。因此，本项目旨在利用网络结构(或拓扑)的敏感度量来开发新的启发式计算方法，以有效地进行网络分析，以应对功能不明确、动态和有噪声的生物网络的复杂性。此外，它的目的是通过利用生物网络数据来帮助理解人类衰老的过程。具体地说，新方法将用于：将关于衰老的知识从模式物种转移到人类，以补充从序列获得的知识；研究动态人类生物网络(通过将当前的静态网络与特定年龄的基因表达数据相结合通过计算获得)，以了解细胞如何随年龄变化；以及对当前网络进行去噪，以产生更高的置信度结果。广泛的影响：理解衰老具有社会重要性。由于网络研究跨越许多领域，所提出的方法将被实施到开源研究软件中，该软件也将作为一种教育工具。将通过新的网络研究课程培训跨学科科学家，进一步促进研究与教育的融合。将向K-12、本科生和研究生提供研究指导，重点是少数民族和女性。将鼓励跨学科合作，以便广泛传播拟议的想法和成果。

The US is growing older because of millions of baby boomers who already started turning 65. Since susceptibility to diseases increases with age, studying molecular causes of aging gains importance. Human lifespan is long, which, in addition to ethical constraints, makes studying human aging difficult. Therefore, aging is studied in simpler ?model? species, e.g., baker?s yeast. Then, the knowledge about aging is transferred from model species to human. Thus far, this transfer has been restricted to genomic sequence comparison, by identifying regions of similarity between sequences of genes in different species (which are believed to be a consequence of functional relationships between the sequences), and by transferring the knowledge from a gene in model species to a sequence-similar gene in human. However, genes (that is, their protein products) carry out biological function by interacting in complex networked ways with one another, instead of acting alone. Hence, it has been argued in the post-genomic era that the wirings among genes in cellular networks could give biological insights over and above sequences of individual genes. Thus, this project hypothesizes that, analogous to genomic sequence research, biological network research will impact our understanding of aging. For example, since not all genes implicated in aging in model species have sequence-similar genes in human, restricting comparison to sequence may limit the transfer of aging-related knowledge to human. Network comparison can help, as it can find regions of similarities between networks of different species and allow for a transfer of the knowledge between such regions.Intellectual merit: Unlike genomic sequence research, biological network research is in its infancy, for the following reasons. Many network problems (including network comparison) are computationally intractable, and hence, efficient approximate (or heuristic) solutions are needed. The function of many genes remains unknown, and hence, it must be discovered from other, better-characterized genes. Even though cells evolve over time, current methods for analyzing systems-level biological networks deal only with their static representations, because dynamic biological network data can not be obtained easily with current biotechnologies, and because there is a lack of efficient methods for dynamic network analysis. Current biological networks are noisy, with many missing and spurious links, due to limitations of biotechnologies as well as human biases during data collection; thus, methods for network de-noising need to be developed. Hence, this project aims to use sensitive measures of network structure (or topology) to develop new heuristic computational methods for efficient network analysis, which can cope with the complexity of functionally uncharacterized, dynamic, and noisy biological networks. Also, it aims to help in understanding the processes of human aging by enabling exploitation of biological network data. Specifically, the new methods will be used to: transfer the knowledge about aging from model species to human to complement the knowledge obtained from sequence; study dynamic human biological networks (obtained computationally by combining current static networks with age-specific gene expression data) to learn about how cells change with age; and de-noise current networks to produce higher-confidence results.Broader impacts: Understanding aging is of societal importance. Since network research spans many domains, the proposed methods will be implemented into open-source research software, which will also serve as an educational tool. Integration of research and education will be promoted further by training interdisciplinary scientists through novel courses on network research. Research supervision will be offered to K-12, undergraduate, and graduate students, focusing on minorities and women. Interdisciplinary collaborations will be encouraged to allow for wide distribution of the proposed ideas and results.