X‑chromosome cfDNA fragment length and methylation entropy as a sex‑specific predictor of mortality via Xi erosion

Hypothesis: Age‑related erosion of X‑chromosome inactivation (Xi) releases a distinct cfDNA signature characterized by shorter fragment lengths and increased methylation entropy that predicts sex‑specific mortality risk independent of circulating sex‑hormone levels.

Rationale: With age, the Xi gains methylation variability and transcriptional reactivation of escapee genes (e.g., KDM6A, FTX) [2,3]. Reactivation loosens chromatin, increases nucleosome turnover, and favors release of ~90‑bp cfDNA fragments from the Xi, while methylated CpGs on the active X remain more protected, yielding longer fragments. Consequently, the proportion of ultra‑short X‑derived cfDNA rises with Xi dysregulation, whereas autosomal cfDNA shows only modest size shifts.

Mechanistic link: Loss of Xi heterochromatin reduces binding of lamina‑associated domain (LAD) proteins, increasing nuclear envelope tension and promoting micro‑nucleus formation that preferentially excises Xi‑rich chromatin. This process releases cfDNA enriched for Xi‑specific SNPs and hypomethylated CpG islands, measurable as a rise in methylation entropy (Shannon entropy of beta values) across X‑linked CpGs.

Prediction: In a longitudinal cohort of men and women aged 60‑80, the ratio of X‑derived cfDNA fragments <100 bp to total cfDNA, combined with X‑linked methylation entropy, will correlate with all‑cause mortality (hazard ratio >1.5 per SD increase) after adjusting for estradiol, testosterone, and traditional risk factors. Conversely, males—who lack a second X—will show a flat relationship because their single X cannot undergo dosage‑compensatory erosion; any increase in X‑cfDNA will reflect generalized nucleolysis rather than Xi‑specific erosion.

Test: Collect plasma, perform whole‑genome cfDNA sequencing, assign reads to X versus autosomes using phased SNPs, compute fragment length histograms and methylation entropy for X‑linked CpGs. Use Cox proportional hazards models. Falsification: If X‑derived cfDNA metrics do not predict mortality beyond autosomal cfDNA or if they equally predict risk in both sexes, the hypothesis is refuted.

Additionally, we would assess whether interventions known to stabilize heterochromatin (e.g., NAD+ boosters, SIRT1 activators) blunt the age‑dependent rise in short X‑cfDNA, providing a mechanistic read‑out for drug efficacy.