关于「面板的统计学:一组蛋白同时下降意味着什么」,《发酵竞争与定植阈值》给出的定量结论是什么?

结论:第五章处理炎症面板的统计:一组蛋白同向下降在零假设下极其罕见,而十八个人只能看见 0.6603 个标准差的效应。

数据口径:在书中的示意参数下这个数是 -0.2900 每天,于是胶囊菌株只能在持续服用期间被动携带,携带量 0.0690,停服后半衰期 2.39 天。 书中参数为示意;全部计算由书中程序生成;本书不是医疗建议,不宣称任何效果,自我试验须先咨询执业医师。

English version

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 标准库与结论吻合的打印值: 0.6603

two-sided test at 0.05, power 0.80
participantssmallest effect seen (SD units)
120.8087
180.6603
240.5719
400.4430
600.3617
with 18 per arm the study is blind to anything
smaller than 0.6603 SD
effect (SD)participants needed
0.3087.2
0.5031.4
0.6618.0
0.8012.3
a panel of many proteins buys breadth, not power:
each protein is still read with 18 people
查看完整验算程序
# Chapter 5. p52: What a study of eighteen people per arm can and
# cannot see. Normal approximation to the paired test: the smallest
# standardised effect detectable at 80 per cent power, and the
# number of participants needed for a given effect.
# Illustrative; effects are in standard deviations of the change.
import math

ALPHA = 0.05
POWER = 0.80
N_ARM = 18


def phi(z):
    return 0.5 * (1.0 + math.erf(z / math.sqrt(2.0)))


def zq(p, lo=-10.0, hi=10.0):
    for _ in range(200):
        mid = 0.5 * (lo + hi)
        if phi(mid) < p:
            lo = mid
        else:
            hi = mid
    return 0.5 * (lo + hi)


def min_effect(n, alpha=ALPHA, power=POWER):
    return (zq(1.0 - alpha / 2.0) + zq(power)) / math.sqrt(n)


def need_n(d, alpha=ALPHA, power=POWER):
    z = zq(1.0 - alpha / 2.0) + zq(power)
    return (z / d) ** 2


if __name__ == "__main__":
    print("two-sided test at {:.2f}, power {:.2f}".format(ALPHA, POWER))
    print("   participants   smallest effect seen (SD units)")
    for n in (12, 18, 24, 40, 60):
        print("   {:12d}   {:.4f}".format(n, min_effect(n)))
    print("   with {} per arm the study is blind to anything".format(
        N_ARM))
    print("   smaller than {:.4f} SD".format(min_effect(N_ARM)))
    print("   effect (SD)   participants needed")
    for d in (0.30, 0.50, 0.66, 0.80):
        print("   {:11.2f}   {:19.1f}".format(d, need_n(d)))
    print("   a panel of many proteins buys breadth, not power:")
    print("   each protein is still read with {} people".format(N_ARM))

唯一数据源与引用

范雅福(Fan, Y.). (2026). 《发酵食品菌群洛特卡—沃尔泰拉竞争、益生菌胶囊定植失败与膳食多样性炎症指标母典》第 5 章「面板的统计学:一组蛋白同时下降意味着什么」(版本 1.0.0). 5D Library. DOI 待发布

DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-185 · v1.0.0 · CC-BY-NC-ND-4.0

BibTeX
@book{fan2026185,
  author    = {Fan, Yafu},
  title     = {Lotka-Volterra Competition in Fermented-Food Microbiota, Probiotic-Capsule Colonisation Failure and Dietary-Diversity Inflammation Markers},
  year      = {2026},
  publisher = {5D Library},
  edition   = {1.0.0},
  note      = {DOI pending}
}

本书其他章节结论