关于「冈珀茨—梅克姆定律:死亡率曲线的形状」,《衰老动力学与毒物兴奋》给出的定量结论是什么?
结论:冈珀茨—梅克姆定律 \(\mu(t) = A + R e^{\alpha t}\) 是对成年人死亡率的稳健经验描述,人类的 MRDT 约为 8 年,但它是典型值而不是常数,估计值还依赖拟合方法。可靠性理论说明,不老化的元件加上冗余的耗尽,就能产生指数型上升与高龄减速,但最简版本只是定性解释;高龄平台是真实现象还是数据误差仍有争论。弗里斯的发病压缩是一个可检验的假说,美国的数据没有支持它;在模型中,只有发病推迟得比死亡更快才会压缩,而要让死亡年龄更集中,必须改变死亡率曲线的斜率。下一章从群体曲线走进细胞:营养感应通路如何在合成与回收之间切换。
数据口径:「自噬」「毒物兴奋」「线粒体低剂量应激」「Zone 2」常被组合成一个故事:吃得太勤把细胞锁在合成状态,适度的饥饿与运动应激开启修复,死亡率曲线因此可以推迟。 本书不是医疗建议,不构成任何抗衰老药物、补充剂或极端禁食的建议,不给任何剂量。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库
| alpha | 0.0866 per year, MRDT = 8.0 years, R = 5.26e-05 |
| age | hazard per year | Makeham share | P(die within the year) | |
| 20 | | 0.00080 | 62.7% | 0.0008 |
| 30 | | 0.00121 | 41.4% | 0.0012 |
| 50 | | 0.00450 | 11.1% | 0.0045 |
| 70 | | 0.02313 | 2.2% | 0.0229 |
| 90 | | 0.12850 | 0.4% | 0.1206 |
| 100 | | 0.30494 | 0.2% | 0.2628 |
| remaining life expectancy: at 30 | 48.7 y, at 50: 30.5 y, at 70: 15.1 y, at 90: 5.2 y |
| fit ages 30-90: MRDT ignoring A | 8.7 y; with A subtracted = 8.0 y (true 8.0) |
| fit ages 50-90: MRDT ignoring A | 8.2 y; with A subtracted = 8.0 y (true 8.0) |
| fit ages 30-60: MRDT ignoring A | 9.7 y; with A subtracted = 8.0 y (true 8.0) |
核心公式
\[\mu(t) = R\,e^{\alpha t},\]
查看完整验算程序
# Gompertz-Makeham hazard mu(t) = A + R exp(alpha t); illustrative: MRDT 8 y, mu(50) ~ 0.004.
# Part 2: refit the MRDT from a synthetic cohort (fixed seed), with and without A.
import math, random
ALPHA = math.log(2) / 8
R = 0.004 / math.exp(ALPHA * 50)
A = 0.0005
mu = lambda t: A + R * math.exp(ALPHA * t)
print(f"alpha = {ALPHA:.4f} per year, MRDT = {math.log(2) / ALPHA:.1f} years, R = {R:.2e}")
print(" age | hazard per year | Makeham share | P(die within the year)")
for t in (20, 30, 50, 70, 90, 100):
print(f" {t:3d} | {mu(t):10.5f} | {A / mu(t):5.1%} | {1 - math.exp(-mu(t)):.4f}")
def remaining(t0, dt=0.01): # remaining life expectancy, integrated to age 120
s, e = 1.0, 0.0
for i in range(int((120 - t0) / dt)):
e += s * dt; s *= math.exp(-mu(t0 + i * dt) * dt)
return e
print("remaining life expectancy: " +
", ".join(f"at {t0}: {remaining(t0):.1f} y" for t0 in (30, 50, 70, 90)))
rng = random.Random(30) # 1,000,000 person-years per year of age
def fit(lo, hi, subtract_A):
xs = list(range(lo, hi + 1))
ys = [math.log(max(rng.gauss(mu(t) * 1e6, math.sqrt(mu(t) * 1e6)) / 1e6
- (A if subtract_A else 0), 1e-7)) for t in xs]
mx, my = sum(xs) / len(xs), sum(ys) / len(ys)
b = sum((x - mx) * (y - my) for x, y in zip(xs, ys)) / sum((x - mx) ** 2 for x in xs)
return math.log(2) / b
for lo, hi in ((30, 90), (50, 90), (30, 60)):
print(f"fit ages {lo}-{hi}: MRDT ignoring A = {fit(lo, hi, False):4.1f} y;"
f" with A subtracted = {fit(lo, hi, True):4.1f} y (true 8.0)")唯一数据源与引用
范雅福(Fan, Y.). (2026). 《衰老动力学、mTOR-AMPK自噬开关与毒物兴奋效应的生物数学模型》第 1 章「冈珀茨—梅克姆定律:死亡率曲线的形状」(版本 1.0.0). 5D Library. https://doi.org/10.5281/zenodo.23770092
DOI: 10.5281/zenodo.23770092 · 阅读本章所在专著 · 5DL-B-030 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026030,
author = {Fan, Yafu},
title = {Ageing Kinetics, the mTOR-AMPK Autophagy Switch and Hormesis},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
doi = {10.5281/zenodo.23770092},
url = {https://doi.org/10.5281/zenodo.23770092}
}