Ma Zhanshan Sam, Mei Jiandong
Computational Biology and Medical Ecology Lab, Kunming Institute of Zoology, Chinese Academy of Sciences, China.
Center for Excellence in Animal Genetics and Evolution, Chinese Academy of Sciences, China.
Comput Struct Biotechnol J. 2022 Mar 30;20:2029-2041. doi: 10.1016/j.csbj.2022.03.027. eCollection 2022.
It is estimated that human body is inhabited by approximately 380 trillions of viruses, which exist in the form of viral communities and are collectively termed as human virome. How virome is assembled and what kind of forces maintain the composition and diversity of viral communities is still an open question. The question is of obvious importance because of its implications to human health and diseases. Here we address the question by harnessing the power of Hubbell's unified neutral theory of biodiversity (UNTB) in terms of three neutral models including standard Hubbell's neutral model (HNM), Sloan's near-neutral model (SNM) and Harris et al. (2017) multi-site neutral model (MSN), further augmented by Ning et al. (2019) normalized stochasticity ratio (NSR) and Hammal et al. (2015) power analysis for the neutral test (PNT). With the five models applied to 179 virome samples, we aim to obtain robust findings given both Type-I and Type-II errors are addressed and possible alternative, non-neutral processes are detected. It was found that stochastic neutral drifts seem prevalent: approximately 65-92% at metacommunity/landscape scales and 67-80% at virus species scale. The non-neutral selection is approximately 26-28% at community scale and 23% at species scale. The false negative rate is about 2-3%, which suggested rather limited confounding effects of non-neutral process on neutrality tests. We postulate that prevalence of neutrality in human virome is likely due to extremely simple structure of viruses (stands of DNA/RNA) and their inter-species homogeneities, forming the foundation of species equivalence-the hallmark of neutral theory.
据估计，人体中存在约380万亿种病毒，它们以病毒群落的形式存在，统称为人类病毒组。病毒组是如何组装的，以及何种力量维持病毒群落的组成和多样性，仍是一个悬而未决的问题。由于其对人类健康和疾病的影响，这个问题显然至关重要。在这里，我们利用哈贝尔的生物多样性统一中性理论（UNTB）的力量，通过三种中性模型来解决这个问题，包括标准的哈贝尔中性模型（HNM）、斯隆的近中性模型（SNM）以及哈里斯等人（2017年）的多地点中性模型（MSN），宁等人（2019年）的标准化随机性比率（NSR）和哈马尔等人（2015年）的中性检验功效分析（PNT）进一步增强了这些模型。将这五个模型应用于179个病毒组样本，我们旨在获得可靠的结果，同时解决I型和II型错误，并检测可能的替代非中性过程。研究发现，随机中性漂移似乎普遍存在：在元群落/景观尺度上约为65 - 92%，在病毒物种尺度上约为67 - 80%。非中性选择在群落尺度上约为26 - 28%，在物种尺度上约为23%。假阴性率约为2 - 3%，这表明非中性过程对中性检验的混杂效应相当有限。我们推测，人类病毒组中中性的普遍存在可能是由于病毒（DNA/RNA链）极其简单的结构及其种间同质性，形成了物种等效性的基础——中性理论的标志。
