Population Health, Max-Planck-Institute for Demographic Research, Rostock, Germany.
Department of Anthropology, New York University, New York, NY, USA.
J Anim Ecol. 2021 Dec;90(12):2901-2914. doi: 10.1111/1365-2656.13590. Epub 2021 Oct 11.
Age-specific fertility trajectories are fundamental to understanding population structure and the evolutionary ecology of diverse life histories. However, characterizing reproductive ageing has been difficult with cross-sectional data, where senescence especially late in life can be confounded by selective disappearance. Addressing such challenge requires longitudinal data tracking the reproductive life span of known individuals, but such data are rare, especially for very long-lived species such as primates. We analyse the entire life span trajectory of annual fertility, from reproductive maturity to death, for 673 free-ranging female rhesus macaques, Macaca mulatta, on Cayo Santiago, Puerto Rico. Using generalized linear mixed-effects models (GLMMs), we first tested if time to death explains the ageing pattern independently of and additionally to chronological age, and if so, whether there is interaction between them. While GLMM captures the patterns in the data well, it is not a generative model. For example, given the GLMM and an individual's reproductive trajectory up to a given age, we cannot directly predict the probability of reproduction or death in the next year. For this reason, we further fitted a hidden Markov chain model (HMM) which allows just such a prediction, and additionally helps infer the process underlying the observed trajectory. We show that, after accounting for individual differences in fertility, reproductive ageing exhibits both age-dependent decline and also an abrupt terminal decline independently of age at death. We infer from the HMM that the underlying process of reproductive trajectory is where individuals cycle between reproductive bouts until they enter an irreversible frail condition that constrains fertility. The findings provide valuable insights into the longitudinal progression of reproductive trajectories in primates, by revealing both age-dependent and age-independent patterns and processes of ageing, and contribute to a growing body of literature on reproductive ageing and senescence across animal taxa.
特定年龄生育轨迹是理解人口结构和不同生命史进化生态学的基础。然而,使用横断面数据来描述生殖衰老具有一定的难度,因为衰老,尤其是生命后期的衰老,可能与选择性消失有关。要解决这个挑战,需要有纵向数据来跟踪已知个体的生殖寿命,但这种数据很少,特别是对于像灵长类这样非常长寿的物种。我们分析了波多黎各卡约圣地亚哥的 673 只自由放养雌性恒河猴(Macaca mulatta)的整个生育期轨迹,从生殖成熟到死亡。使用广义线性混合效应模型(GLMM),我们首先测试了死亡时间是否独立于年龄,并且是否与年龄存在相互作用,从而解释衰老模式。虽然 GLMM 很好地捕捉了数据中的模式,但它不是一个生成模型。例如,给定 GLMM 和个体在特定年龄之前的生殖轨迹,我们不能直接预测下一年的繁殖或死亡概率。出于这个原因,我们进一步拟合了一个隐马尔可夫链模型(HMM),它不仅可以进行这样的预测,而且还可以帮助推断观察到的轨迹背后的过程。我们表明,在考虑个体生育能力差异后,生殖衰老表现出年龄依赖性下降和与死亡年龄无关的突然终端下降。我们从 HMM 推断,生殖轨迹的潜在过程是个体在繁殖之间循环,直到他们进入一个不可逆转的脆弱状态,限制了生育能力。这些发现为了解灵长类动物生殖轨迹的纵向进展提供了有价值的见解,揭示了衰老的年龄依赖性和非年龄依赖性模式和过程,并为跨动物类群的生殖衰老和衰老的不断增加的文献做出了贡献。