Finding. Human sperm are among the most specialized cells we make, and we still argue where their proteins live, who secreted them, and why one donation differs from the next. van der Hoeven, Olsen and colleagues report 10,828 proteins across six donors, the deepest sperm proteome in the stack. Abundances agree between men. What holds the map together is a conserved core of mitochondrial energy-metabolism proteins. A smaller slice changes between donations from the same man. They mechanically split heads from tails and assigned thousands of previously homeless proteins to a compartment. They compared seminal fluid from vasectomized and non-vasectomized men to separate sperm- and epididymis-derived proteins from accessory-gland secretions. A single-cell workflow recovers up to about 1,000 protein groups per spermatozoon. In a total fertilization failure cohort, the recurrent problem is mitochondrial, not the sperm-egg fusion set.
Why this paper matters
In vitro fertilisation (IVF) programs still meet couples whose sperm look adequate and whose eggs fail to fertilize. The reflex is fusion genetics. This atlas says look at the midpiece first. A conserved mitochondrial core that is stable across men is also the thing that breaks when fertilization fails completely.
It is also a methods object. Head/tail splits, vasectomy contrasts, and single-sperm depth give every later sperm paper a localization grammar.
How to read the score
High eighties. Human tissue, unusual depth, a mitochondrial thesis that is not a keyword dump, and a clinical cohort with a directional claim. Confidence is high for the atlas size and the conserved-core sentence. It is medium for the fertilization-failure mechanism until we see the protein list.
Caveats
Six donors. Total fertilization failure n is not in the abstract. Single-cell coverage is an order of magnitude below the bulk atlas. This is not a diagnostic test. Do not tell a clinic to skip fusion-gene work from this brief.
What to do with it
If you work male infertility, pull the mitochondrial energy-metabolism panel before you write "fusion defect." If you build sperm proteomes, cite 10,828 and the head/tail map. If you do single-cell proteomics, treat 1,000 protein groups as the current per-cell ceiling they claim, not the bulk depth.