Mendelian randomization works by borrowing a coin flip that nature already made. The genetic variants setting a person's average tyrosine slightly higher or lower were dealt before diet or illness could confound anything, which is why the University of Hong Kong and University of Georgia team leaned on that design rather than on the correlation alone [2][6]. Sick people metabolise amino acids differently, and an association between a blood metabolite and dying is exactly the kind of finding that reverses direction under scrutiny.
The design is the interesting part, and it is also where the reporting thins. The paper in Aging-US is titled for both amino acids [3], but the summary from Impact Journals gives the male figure as nearly one year without saying what increment of tyrosine it is per, and without a test of whether the male and female estimates differ from one another [16]. Two subgroup results sitting side by side do not amount to a formal interaction test; the sex-specific story rests entirely on the contrast between them.
Take both estimates at face value and they bound their own reach. About a year of penalty in men against nothing detectable in women leaves tyrosine able to account for, at most, roughly a year of the difference in how long men and women live [17].
The correlational and genetic analyses were run in the same participants [1][5], so the second draws on the same sample as the first, a repeat test rather than an independent replication. It is a different question put to the same people, and it inherits that cohort's diet and ancestry.
A null result in women marks the limit of what the study could detect; a true effect of zero remains an open claim, not an established one. The release reports no significant effect in women and does not report how precisely that null was estimated [9][16]. What keeps the male result from being easy to dismiss is that the biology is coherent. Tyrosine is the precursor the body uses to make dopamine [13], which is precisely why it is marketed for focus, and a molecule feeding a signalling pathway is a reasonable place to look for consequences beyond protein synthesis.
