HIV Incidence Assay via Deep Sequencing and Statistical Tests

描述(申请人提供)：我们最近的研究开发了急性序列进化模型，该模型已用于解释来自102名急性HIV B亚型感染者、69名急性HIV C亚型感染者以及大量SIV感染猕猴的序列克隆。我们的模型能够评估源自单一传播病毒株的艾滋病毒感染的序列多样性，并估计感染时间，这对艾滋病毒研究界做出了重大而有益的贡献。我们建议扩大我们对宿主内HIV多样化建模的研究活动，以开发一种新的检测方法来识别新的、最近的HIV感染。在艾滋病毒/艾滋病预防领域，至关重要的是评估某一地区最近有多少人受到感染，以评估预防和干预试验的可行性。目前，从血液样本中诊断艾滋病毒感染是可能的，但还没有开发出可靠的分析方法来预测一个人感染了多久。假设抗体滴度和亲和力随着时间的推移而增加，已经使用了各种失谐抗体分析，或基于亲和力的评估。然而，血清学分析是有问题的，因为(I)抗体反应的成熟速度在不同的人之间是不同的，(Ii)一些CD4计数低或病毒载量低的受试者可能被错误地计算为最近感染，以及(Iii)如果感染的病毒分支与分析中使用的抗原分支不同，则血清学分析可能会受到负面影响(这在混合分支流行病中是一个特别的问题)。这项建议旨在为发明一种新的检测方法提供经验和理论基础，该方法可以区分新的、偶发感染和慢性、无症状感染。该提案的主要创新包括：(I)它全面集成了下一代超深层焦磷酸测序数据和生物数学模型，以及(Ii)一种新的统计设计，估计了创始人病毒的数量和感染持续时间。我们建议的研究将集中于克服发展基于测序的分析的两个主要障碍。首先，该分析可能会将具有多个不同创始人菌株的早期感染错误归类为慢性感染。其次，化验是否能区分感染后期获得的慢性样本和偶发样本，还值得商榷。我们建议从不同流行区和不同感染阶段的不同人群中收集大规模的HIV序列数据。超深测序数据与新的统计设计相结合，将引导我们发明一种能够识别偶发感染的新方法。拟议的工作将根据宿主内艾滋病毒多样化的特点提供可靠的检测方法，从而推动艾滋病毒预防研究。

DESCRIPTION (provided by applicant): Our recent studies developed acute sequence evolution model which has been applied to interpret sequence clones derived from 102 subjects with acute HIV subtype B infection, from 69 subjects with acute HIV subtype C infection, and from numerous SIV-infected macaques. Our model enabled an assessment of the sequence diversity in HIV infections originating from a single transmitted viral strain and the estimation on the period of infection, which has been a significant and beneficial contribution to HIV research community. We propose to expand our research activity of modeling intrahost HIV diversification toward the development of a novel assay identifying new, recent HIV infections. In the HIV/AIDS prevention field, it is critical to assess how many people have been recently infected in a given area in order to evaluate the feasibility of prevention and intervention trials. The diagnosis of HIV infection is currently possible from blood samples but reliable assays predicting how long an individual has been infected have not been developed. Various detuned antibody assays, or avidity-based assessments, have been used, assuming that antibody titer and avidity increases with time. However, serologic assays are problematic because (i) the rate of maturation of the antibody response varies between different individuals, (ii) some subjects with low CD4 counts or low virus loads may be inaccurately counted as having recent infection, and (iii) serologic assays can be negatively affected if the infecting virus clade differs from the clade of the antigen used in the assay (this is a particular problem in mixed-clade epidemics). This proposal aims to provide empirical and theoretical foundations for inventing a novel assay that distinguishes new, incident infections from chronic, asymptomatic infections. Major innovations of the proposal include (i) its comprehensive integration of next-generation ultradeep pyrosequencing data and biomathematical modeling and (ii) a novel statistical design estimating the number of founder viruses and the duration of infection. Our proposed study will focus on overcoming two primary barriers to the development of a sequencing based assay. First, the assay potentially mis-classifies early infections with multiple distinct founder strains as chronic infections. Second, it is questionable whether the assay can distinguish incident samples from chronic ones obtained at late stages of infections. We propose to collect a large scale of HIV sequence data from diverse population in different epidemic regions and different stages of infections. Ultradeep sequencing data in conjunction with novel statistical designs will lead us to invent a novel assay that is capable of identifying incident infections. The proposed work will advance HIV prevention research by providing a reliable assay based on the characteristics of intrahost HIV diversification.