Evidence level: B
要证明人身上有一个 29.5 天的表达节律,唯一现实的入口是外周血。而这条路最难的一步,是把 24 小时的摆动从 29.5 天里剥出去。
正文
中医把”月满则血气积”当常识讲了两千年。上一篇《人类近月钟的分子实体》给出了一张”照虫找钟”的路线图:海洋环节虫的近月钟已解到分子层面——昼夜钟与近月钟可以被独立打掉(PMID 24075994),近月相位有明确的转录组-蛋白质组签名(PMID 30767890),摇蚊的昼夜与近月计时适应有清晰的基因组基础(PMID 27871090)。接下来的问题很直接:在人类身上,这件事能不能测? 现实的入口只有一个——外周血。而这条路最难的部分不在采血,在采样设计。
外周血是唯一能反复读的器官
要证明一个 29.5 天的表达成分存在,必须对同一个人反复采样多个月,能做这件事的只有外周血。基础是有的:人 PBMC 中钟基因的表达模式已被系统刻画(PMID 16096739 报告了非典型模式,PMID 18336944 给出 10 个钟基因在 PBMC 中的表达谱)。”读得到”这一环不是障碍。
障碍是:你读到的从来不只是钟。 外周血钟基因表达极易被状态推动——轮班护士的白细胞钟基因表达随排班漂移(PMID 22552736、PMID 24402410,后者同时测到尿褪黑素与血清雌二醇的同步变化);一次体内内毒素暴露就能同步并压制人白细胞钟基因表达(PMID 20081528);库欣病患者外周钟系统明确异常(PMID 32354240),内源性高皮质醇症会破坏免疫细胞的昼夜振荡(PMID 39612654);连造血干细胞动员这种治疗性干预都会改写外周血钟基因表达(PMID 26158633)。外周血不是一支干净的体温计,它更像一间同时开着一二十个振动源的房间。
昼夜必须让开,月相才露得出来
现有的周期成分检测工具——JTK_CYCLE(PMID 20876817)、RAIN(PMID 25326247)、基于自回归谱估计的谐波回归(PMID 20529902)、似然比检验(PMID 34117739),乃至更新的贝叶斯全基因组框架(PMID 41148221)——都围绕昼夜尺度、密集采样发展。它们有一个共同前提:采样间隔远小于目标周期的四分之一。
把这套工具直接搬到 29.5 天周期上会撞上硬约束。若一天采一次(外周血最现实的节奏),对 29.5 天来说采样点在相位上几乎冻结:连采 100 天,写下的其实是 100 个”几乎同一相位”的读数,周期再长也检不出来。若只在几个相位点采(新月/上弦/满月/下弦),采样密度又太低,检测功效趋近于零。
于是设计被迫两难:为了让昼夜让开,每次采样要固定在同一个昼夜时刻;为了覆盖月相,采样又必须散在多个相位。 两个要求在真实研究里互相拉扯,不少”月相效应”就是这样被做成了昼夜效应的影子,或者一次采样时机的偶然。
还有两个更深的坑
一、28 天和 29.5 天在几个周期内分不开。 女性受试者身上同时跑着月经周期与月相周期,周期只差约 5%。只观察三四个周期时,两者在数学上几乎不可分离;要靠相位漂移拆开,至少得观察一年以上(两周期约每 14 个月重新对齐一次)。PMID 38598621 认为卵巢周期由内源月级计时系统驱动,这本身就说明两个尺度在生理上需要分开讨论。性别还会改变白细胞对刺激的表达反应(PMID 27832807)。这不是”顺手校正一下”能解决的,是必须写进设计的前提。
二、转录层可能不是对的那一层。 环节虫的硬证据是:近月钟的调控主要发生在蛋白质层,RNA 变化小、蛋白变化强(PMID 30767890)。人类这边同样不乐观:mRNA 与蛋白的相关性天生有限(PMID 32916130 的人体定量蛋白质组图展示了这种不一致的普遍性)。理论上该做转录组加蛋白质组双平台,但血浆蛋白质组本身就有显著昼夜波动(PMID 40846907、PMID 31875183),等于把同一份混杂在另一个平台上再走一遍。
效应量该有多大的期待
一个必须提前记下的数字:月相对人类行为的影响,本身就处在”有报告、复制不稳”的状态——那项发现月相与睡眠关联的研究(PMID 23891110)没有被更大样本的人群研究复制(PMID 26498230),另一项只看到性别交互(PMID 26096730)。行为层的效应还在检出边界上晃,指望分子层的表达成分幅度更大,并不现实。
所以对”人类外周血有没有 29.5 天表达成分”这个问题,目前最诚实的回答是:没人测过,而且它比看起来难得多。真正需要的不是更灵敏的算法,而是一套愿意花一到两年、固定昼夜时刻、覆盖多个相位、分层性别、双平台采样的纵向方案。
《素问·六节藏象论》说”五日谓之候,三候谓之气,六气谓之时”。古人把观察的最小单位定在”五日”,这不是随手定的,而是在承认:分辨率有代价,你得先决定一”候”多长,才谈得上看不看得见。 现代分子检验在做同一件事——把一个 29.5 天的周期切成若干采样点,再问这一点够不够密。还有一个更根本的差别:中医问的是”月满时人是不是更实”,这是状态问句;分子检验把它翻译成”某通路存在 29.5 天周期的表达成分”,这是归属问句。前者承认你只能看到人此刻的样子,后者要求你指出是哪一个零件在报时。 古人没有能力问后者,我们也还没能力回答前者。这中间的空白,就是接下来要填的那一格。
证据与局限
已验证: 环节虫近月钟的分子通路完整——昼夜钟与近月钟可被独立解耦(PMID 24075994)、近月相位存在转录组与蛋白质组签名且信号主要在蛋白质层(PMID 30767890)、近月计时适应的基因组基础(PMID 27871090);人外周血 PBMC 中昼夜钟基因表达可测且模式已被系统刻画(PMID 16096739、PMID 18336944);外周血钟基因表达受状态强烈影响——轮班(PMID 22552736、PMID 24402410)、内毒素(PMID 20081528)、皮质醇过量(PMID 32354240、PMID 39612654)、造血动员(PMID 26158633);周期成分检测算法成熟(PMID 20876817、PMID 25326247、PMID 20529902、PMID 34117739、PMID 41148221、PMID 25793520);人体 mRNA 与蛋白水平不一致普遍存在(PMID 32916130);血浆蛋白质组存在昼夜波动(PMID 40846907、PMID 31875183);女性卵巢周期由内源月级计时系统驱动(PMID 38598621);白细胞表达反应存在性别差异(PMID 27832807)。
部分验证: 人类月相-睡眠关联存在但复制不稳(PMID 23891110 阳性、PMID 26498230 阴性、PMID 26096730 仅性别交互)——提示即使行为层效应也处于检出边界,分子层效应量不宜乐观预期。
待验证: ①人类外周血 CRY/PER/CLOCK 类基因是否存在独立于昼夜相位的 29.5 天表达成分(尚无研究执行);②在稀疏采样与稀疏相位覆盖条件下,周期检测算法对 29.5 天成分的功效与假阳率尚无标定;③月经周期成分与月相成分在同一批外周血样本中的可分离性(需要多长观察窗才能靠相位漂移拆开);④蛋白质组层面的月相成分是否强于转录层(环节虫提示如此,人类未测)。
相关阅读
- 人类近月钟的分子实体:对照环节虫通路扫描哺乳动物同源基因的月相表达
- 人体内源近月振荡器:月亮时钟是否刻在基因里
- 近月钟靶蛋白的哺乳动物同源扫描
- 0.3 勒克斯落在两个人身上是两个剂量:暗适应时长如何重标定褪黑素抑制阈值
- 研究池:生物频率密码
证据等级: B
参考文献: PMID 24075994, PMID 30767890, PMID 27871090, PMID 16096739, PMID 18336944, PMID 22552736, PMID 24402410, PMID 20081528, PMID 32354240, PMID 39612654, PMID 26158633, PMID 20876817, PMID 25326247, PMID 20529902, PMID 34117739, PMID 41148221, PMID 25793520, PMID 32916130, PMID 40846907, PMID 31875183, PMID 38598621, PMID 27832807, PMID 23891110, PMID 26498230, PMID 26096730
免责说明:本文属传统知识与科学文献的整理、对照与研究设计讨论,仅供理解与参考,不构成医疗建议,不能替代专业诊断与治疗。文中涉及的效应量、频率区间与暴露结论多为研究报告层面的推断,个体情况差异很大;如有身体不适或健康疑虑,请就医并由专业医师判断。
Finding a 29.5-day clock in human blood: The challenge isn’t drawing blood, but making day and night yield
Evidence level: B
To prove the existence of a 29.5-day expression rhythm in the human body, the only realistic entry point is peripheral blood. The most difficult step on this path is not the act of phlebotomy itself, but designing a sampling protocol that allows the 24-hour cycle to yield to the lunar phase.
Main Text
Traditional Chinese medicine has treated “when the moon is full, blood and qi accumulate” as common knowledge for two millennia. The previous article, “The Molecular Entity of the Human Near-Lunar Clock,” provided a roadmap for “finding the clock by looking at the insect”: the near-lunar clock in marine annelids has been resolved to the molecular level—the circadian and near-lunar clocks can be independently disrupted (PMID 24075994), near-lunar phases have distinct transcriptomic and proteomic signatures (PMID 30767890), and the genomic basis for circadian and near-lunar timekeeping adaptation in mosquitoes is clear (PMID 27871090). The next question is straightforward: Can this be measured in humans? The only realistic entry point is peripheral blood. The most difficult part of this path is not the sampling itself, but the sampling design.
Peripheral Blood Is the Only Organ That Can Be Read Repeatedly
To prove the existence of a 29.5-day expression component, one must sample the same individual multiple times over several months; only peripheral blood allows this. The foundation exists: clock gene expression patterns in human PBMCs have been systematically characterized (PMID 16096739 reported an atypical pattern, PMID 18336944 provided expression profiles of ten clock genes in PBMCs). The ability to “read” this signal is not an obstacle.
The obstacle is: What you read is never just the clock. Peripheral blood clock gene expression is easily driven by physiological state—shift nurse leukocyte clock gene expression drifts according to shift schedules (PMID 22552736, PMID 24402410, the latter simultaneously measuring synchronized changes in urinary melatonin and serum estradiol); a single bout of endotoxin exposure can simultaneously suppress human leukocyte clock gene expression (PMID 20081528); patients with Cushing’s syndrome show clear abnormalities in the peripheral clock system (PMID 32354240), and endogenous hypercortisolism disrupts circadian oscillations in immune cells (PMID 39612654); even therapeutic interventions like hematopoietic stem cell mobilization rewrite peripheral blood clock gene expression (PMID 26158633). Peripheral blood is not a clean thermometer; it is more like a room with twenty or more vibrating sources turned on simultaneously.
Day and Night Must Yield for the Lunar Phase to Appear
Existing tools for detecting periodic components—JTK_CYCLE (PMID 20876817), RAIN (PMID 25326247), harmonic regression based on autoregressive spectral estimation (PMID 20529902), likelihood ratio tests (PMID 34117739), and newer Bayesian whole-genome frameworks (PMID 41148221)—have all developed around circadian scales and dense sampling. They share a common premise: the sampling interval must be far smaller than one-quarter of the target cycle.
Applying these tools directly to a 29.5-day cycle hits a hard constraint. If sampling occurs once per day (the most realistic rhythm for peripheral blood), the sampling points for a 29.5-day cycle are almost frozen in phase: even collecting over 100 days yields readings that are essentially “almost the same phase,” making detection of longer cycles impossible. If sampling occurs only at specific phases (new moon, first quarter, full moon, last quarter), the sampling density is too low, driving detection power toward zero.
This forces a design dilemma: To allow day and night to yield, each sample must be fixed at the same circadian time; to cover lunar phases, sampling must be scattered across multiple phases. These two requirements pull against each other in real-world research, causing many “lunar phase effects” to be mistaken for circadian artifacts or random sampling timing coincidences.
Two Deeper Pits Remain
1. 28 Days and 29.5 Days Are Indistinguishable Over Several Cycles. Female subjects simultaneously run menstrual cycles and lunar cycles, differing by only about 5%. Observing just three or four cycles makes them mathematically nearly inseparable; separating them via phase drift requires observation for at least one year (two cycles realign approximately every 14 months). PMID 38598621 posits that the ovarian cycle is driven by an endogenous monthly timing system, which itself indicates that these two scales require separate physiological discussion. Gender also alters leukocyte expression responses to stimuli (PMID 27832807). This is not something that can be “fixed with a simple correction”; it must be a prerequisite written into the study design.
2. The Transcriptome May Not Be the Correct Layer. The hard evidence in annelids is that near-lunar clock regulation occurs primarily at the protein level, with small RNA changes but strong protein changes (PMID 30767890). The situation in humans is similarly pessimistic: the correlation between mRNA and protein is inherently limited (PMID 32916130’s human quantitative proteomics study demonstrates the universality of this inconsistency). Ideally, one should use dual platforms combining transcriptomics and proteomics; however, the plasma proteome itself exhibits significant circadian fluctuations (PMID 40846907, PMID 31875183), effectively running the same confounding signal through another platform.
Expectations for Effect Size
A number that must be noted in advance: the effect of the lunar phase on human behavior is already in a state of “reported but unstable replication”—the study linking lunar phase to sleep (PMID 23891110) was not replicated by larger population studies (PMID 26498230), while another found only a gender interaction (PMID 26096730). Effects at the behavioral level hover near the detection threshold; expecting larger magnitude effects at the molecular expression level is unrealistic.
Therefore, for the question of whether “human peripheral blood has a 29.5-day expression component,” the most honest current answer is: No one has measured it, and it is far harder than it appears. What is truly needed is not more sensitive algorithms, but a longitudinal protocol willing to spend one to two years, fixed at circadian times, covering multiple phases, stratified by gender, and utilizing dual platforms.
The Su Wen · Six Sections of Hidden Images states: “Five days are called a ‘hou’, three hou are called a ‘qi’, six qi are called a ‘shi’.” Ancient observers set the minimum unit of observation at “five days.” This was not arbitrary; it acknowledges: Resolution has a cost. You must first decide the length of one ‘hou’ before you can even speak of seeing or not seeing. Modern molecular testing does the same thing—it slices a 29.5-day cycle into sampling points and asks if that density is sufficient. There is also a more fundamental difference: Traditional medicine asks, “Is the human body more ‘solid’ at the full moon?” This is a state question; molecular testing translates it to “Does a specific pathway have a 29.5-day periodic expression component?” This is an attribution question. The former acknowledges you can only see how the person looks at this moment; the latter demands you identify which specific part is signaling the time. The ancients lacked the ability to ask the latter, and we still lack the ability to answer the former. The gap between them is the blank space to be filled next.
Evidence and Limitations
Verified: The molecular pathways of the annelid near-lunar clock are complete—the circadian and near-lunar clocks can be independently uncoupled (PMID 24075994), near-lunar phases have transcriptomic and proteomic signatures with signals primarily at the protein level (PMID 30767890), and the genomic basis for near-lunar timekeeping adaptation is established (PMID 27871090); circadian clock gene expression in human peripheral blood PBMCs is measurable and systematically characterized (PMID 16096739, PMID 18336944); peripheral blood clock gene expression is strongly influenced by state—shift work (PMID 22552736, PMID 24402410), endotoxin (PMID 20081528), excess cortisol (PMID 32354240, PMID 39612654), and hematopoietic mobilization (PMID 26158633); algorithms for detecting periodic components are mature (PMID 20876817, PMID 25326247, PMID 20529902, PMID 34117739, PMID 41148221, PMID 25793520); human mRNA and protein levels are frequently inconsistent (PMID 32916130); the plasma proteome exhibits circadian fluctuations (PMID 40846907, PMID 31875183); the female ovarian cycle is driven by an endogenous monthly timing system (PMID 38598621); leukocyte expression responses show gender differences (PMID 27832807).
Partially Verified: Human lunar phase-sleep associations exist but show unstable replication (PMID 23891110 positive, PMID 26498230 negative, PMID 26096730 gender interaction only)—suggesting that even behavioral effects may lie at the detection threshold, making optimistic expectations for molecular effect sizes inappropriate.
To Be Verified: ① Whether CRY/PER/CLOCK-like genes in human peripheral blood possess a 29.5-day expression component independent of the circadian phase (no studies have addressed this yet); ② Under conditions of sparse sampling and sparse phase coverage, the power and false-positive rates of periodicity detection algorithms for a 29.5-day component remain uncalibrated; ③ The separability of menstrual cycle components and lunar phase components within the same batch of peripheral blood samples (how long an observation window is required to separate them via phase drift); ④ Whether lunar phase components are stronger at the proteome level than the transcriptome level (annelid evidence suggests yes; humans remain unmeasured).
Related Reading
- The Molecular Entity of the Human Near-Lunar Clock: Mapping Lunar Expression Across Mammalian Homologs of Annelid Pathways
- Endogenous Near-Lunar Oscillators in Humans: Is the Moon Clock Carved into Our Genes?
- Scanning for Mammalian Homologs of Near-Lunar Clock Target Proteins
- 0.3 Lux Falls on Two People as Two Doses: How to Recalibrate Melatonin Suppression Thresholds via Dark Adaptation Duration
- Research Pool: Biological Frequency Codes
Evidence Level: B
References: PMID 24075994, PMID 30767890, PMID 27871090, PMID 16096739, PMID 18336944, PMID 22552736, PMID 24402410, PMID 20081528, PMID 32354240, PMID 39612654, PMID 26158633, PMID 20876817, PMID 25326247, PMID 20529902, PMID 34117739, PMID 41148221, PMID 25793520, PMID 32916130, PMID 40846907, PMID 31875183, PMID 38598621, PMID 27832807, PMID 23891110, PMID 26498230, PMID 26096730
Disclaimer: This article compiles traditional knowledge and scientific literature for comparison and research-design discussion only. It is not medical advice and cannot replace professional diagnosis or treatment. Effect sizes, frequency ranges and exposure conclusions cited here are largely inferences at the research-report level, and individual circumstances vary widely. If you have any discomfort or health concerns, please consult a physician.
Leave a Reply