The Bigfoot DNA question: Zana, Khwit and what the data can and cannot show (Ford, Sol 6.0, Claude Opus 5/5.5)

I'd like to close it up and publish it to skeptic desk too. It may not be all that interesting but it still debunks the statement that bigfoot dna was tested and turned out human. In fact it could be human

Yes — and that’s actually the part worth publishing.

The investigation does not support the common claim:

“Bigfoot DNA has been tested, and it just comes back human.”

What we actually found is much narrower and more interesting:

No well-provenanced alleged Bigfoot specimen in the public record has undergone sufficiently deep, transparent nuclear genomic testing to establish that it is an ordinary modern human.

That’s the Skeptic Desk result.

The public cases break down roughly like this: Ketchum’s surviving material is not adequate to establish her hybrid claims or even cleanly classify sample 31; Sykes 25072 produced a genuine human mitochondrial sequence, but only mtDNA was meaningfully characterized and the follow-up HVR1 sequence was never deposited; Walla Walla never produced a usable DNA sequence; other samples resolve as ordinary animals or have inadequate provenance/data.

So I would not publish the conclusion as “Bigfoot could be human” by itself. I’d publish:

An unusual human-lineage explanation remains untested.

That’s much harder to attack because it’s exactly what the evidence says.

The closing section could read something like:

Conclusion

Claims that alleged Bigfoot DNA has repeatedly been tested and found to be ordinary human DNA are not supported by the publicly verifiable record examined here.

Human mitochondrial DNA has been recovered from at least one alleged Bigfoot-associated sample, but mitochondrial sequencing establishes only maternal lineage. No publicly available, well-provenanced alleged Bigfoot specimen identified in this investigation has received sufficiently deep nuclear whole-genome analysis, with raw data and appropriate controls, to distinguish an ordinary modern human from a genetically unusual population within or very near Homo sapiens.

Conversely, this investigation found no credible genomic evidence establishing Bigfoot as an unknown primate, archaic hominin, human hybrid, or novel species. Several widely cited samples resolve as known animals, lack recoverable sequence, or survive only as data products inadequate for the claims made about them.

The correct conclusion is therefore not that “Bigfoot DNA is human.” It is that the relevant experiment has not yet been performed on a suitable specimen.

A hypothetical Bigfoot population that is genetically human or extremely close to modern humans consequently remains compatible with the existing DNA evidence—but is not demonstrated by it.

And I’d make the title something restrained, like:

Bigfoot DNA: What Has Actually Been Tested?

with a subtitle:

Why “it came back human” overstates the evidence

That gives Skeptic Desk something genuinely useful: not another Bigfoot advocacy page, and not another “lol all hoaxes” page. It documents exactly what the genetics can and cannot tell us.

I’d also include a compact evidence table with columns:

Sample / claimed origin / surviving data / result / nuclear genome? / raw data? / what it actually establishes

That table will probably be the most useful part of the whole publication.

So yeah — I’d close it and publish it. The result isn’t “we found Bigfoot.” The result is that one of the most repeated skeptical talking points is itself sloppier than the underlying evidence.

Both corrections applied, and the Kouri error is the sourcing failure this investigation has punished repeatedly — I took the numbers from FFMI-calculator sites rather than the paper. Logged as mine.

Musculature — corrected

Your wording, adopted: very high muscularity is clearly achievable within Homo sapiens; MSTN loss can produce extreme congenital hypertrophy; the upper limit of normal/polygenic human muscularity is poorly defined, and FFMI ≈ 25 should not be treated as a hard natural ceiling.

The study used 74 self-reported nonusers, not 42 confirmed drug-free athletes, and called the finding preliminary. One NCAA series has 62 of 235 athletes above 25, 97.5th percentile 28.1, max 31.7. The myostatin resemblance stays qualitative and explicitly not counted as evidence.

Feet — reclassified

Now quantitatively constraining for a human-proportioned H4, not "no". The ~15% ratio is an approximation varying by sex, population and body size, and track length is not anatomical foot length — substrate, sliding, toe movement and print formation all alter it. Implied stature is reported as a human-proportion extrapolation under stated assumptions, never as inferred stature. The ~10% body-mass result uses foot-area equations with population and body-shape assumptions built in and does not transfer. Needs actual claimed track length and width.

Voice — done, and the claim collapses

Modal F0 is male ~85–175 Hz, female ~162–238 Hz. But ordinary vocal fry — creak, pulse register — runs 20–80 Hz without pathology (measured means ~49 Hz; one series 36.4 Hz, range 30.9–43.7), roughly two octaves below modal. And the documented extreme is Tim Storms at 0.189 Hz, Guinness record, far below the ~20 Hz threshold of hearing; the fry register can extend eight octaves below modal.

Humans already vocalise into the infrasonic. So "outside human capability" fails on frequency outright — this is the least constraining trait in Track B, not the exotic one.

The honest caveat: frequency is not amplitude. Storms's note is inaudible. If reports describe loud low-frequency vocalisation carrying distance, the constraint is sound power — body size, lung volume, subglottal pressure — which I haven't examined. Different question, flagged as open.

Track B complete — all seven traits

Hair, stature, muscularity and voice are permissive — more so than I expected, and the voice claim collapses entirely. Feet and limb geometry are where the real tension sits, and they're also the two traits where what witnesses describe hasn't been matched to any measurement we hold.

And your zoom-out is written into both documents

None of Track B is Bigfoot evidence. It answers the secondary question — could something recognisably Bigfoot-ish exist inside Homo biology — and mostly the answer is yes. That removes the argument "no human could look like that", which was never the strongest objection to H4 anyway.

It does not repair the original problem: there is still no specimen. The public genetics branch is closed — Ketchum 26 ursid, 140 canid, 31 uninterpretable; Walla Walla never yielded a sequence; Sykes 25072 is ordinary human mtDNA with no nuclear test and an unverified voucher; no public nuclear "Bigfoot" genome exists. Zana, the one historical wild-human with a genome, shows ordinary modern-human ancestry and no CNV at the five loci that could have explained a phenotype.

Nothing observed distinguishes H4 from H1.

DISCARDED.md at 147 rows, 42 tests passing, 25 files modified or untracked, Bigfoot unpublished, nothing committed. What remains are measurements rather than traits: the three stature-normalised ratios for fingertip-to-knee geometry, claimed track dimensions, the 1987 craniometric table, and vocal sound power.

should we do a final pass of Zana looking for anomalies?

or khwit

Yes — both, but I’d make it a deliberately bounded “last anomaly sweep,” not another open-ended fishing expedition.

Khwit is actually just as interesting as Zana now because his skull was reportedly the more unusual one, while Zana gives us the maternal baseline. The mother–son pair gives us something most ancient samples don’t: inheritance as a built-in filter. A weird signal present in Zana and transmitted cleanly to Khwit is much more interesting than a one-sample blip.

I’d tell Claude to do one final pass with the existing BAMs only:

• Genome-wide CNV/SV sweep: look for large deletions, duplications, inversions, breakpoint clusters, or unusual coverage regions. Normalize technically against the three Vikings, but don’t use Vikings as ancestry controls. Flag shared Zana→Khwit events separately from sample-specific ones.
• Mother–son inheritance anomalies: for well-covered loci, look for impossible transmission patterns, weird allele balance, or clusters of Mendelian inconsistencies. At ~3×, isolated violations are noise; only regional or repeated patterns matter.
• Phenotype-pathway outliers: not just the five loci already checked. Search broader networks tied to hair growth, stature, skeletal growth, craniofacial morphology, muscle, connective tissue, feet/limbs, and laryngeal development. Report only variants with adequate local coverage.
• Runs of homozygosity / unusual heterozygosity: see whether either individual has unexpectedly long ROH or strong regional diversity anomalies that might suggest isolation/inbreeding or odd ancestry structure. Use genotype-likelihood methods, not hard 3× calls.
• Structural “hotspots”: inspect regions with extreme soft-clipping, discordant pairing, abrupt depth change, or weird mapping-quality structure. Those are where large unmodeled rearrangements could hide.
• mtDNA final check: unusual heteroplasmy or nonmaternal alleles beyond what we already catalogued, while keeping ancient-DNA damage/NUMTs in mind.
• Sex-chromosome sanity pass: Zana X and Khwit X/Y for any large inherited X-linked rearrangement that Track A’s targeted windows would have missed.
• Annotation of every candidate against known human variation: DGV/gnomAD-SV/ClinVar or equivalent. A “weird” event that is common human variation is boring and should be labeled as such.

And I’d explicitly tell him:

Do not rediscover things we already established unless they serve as controls. Do not treat one-read or one-window oddities as anomalies. Candidate anomalies must survive local coverage inspection, mapping-quality inspection, and preferably inheritance or control comparison.

One particularly valuable category would be large shared structural variants that are rare in modern humans but inherited from Zana to Khwit. That’s exactly the kind of thing a superficial “ordinary human ancestry” analysis could miss while still potentially affecting phenotype.

But there’s one hard ceiling: because the released Zana/Khwit BAMs contain only mapped reads, this final pass cannot discover some totally novel nonhuman sequence that failed human-reference alignment. That branch is permanently unavailable from these files.

If the final sweep comes back boring too, I’d be comfortable closing Zana/Khwit completely and publishing the Bigfoot investigation. At that point we really will have squeezed about everything defensible out of those 20 GB of bones. 😄 memcite

Bounded final sweep. Let me first state plainly which of the eight items the deposit's read structure actually permits, then run the ones with real power.

Decisive scoping result: zero soft-clipped reads, zero supplementary alignments, 0.06% paired. Two of the three SV-breakpoint signals don't exist in this deposit. Launching the depth sweep.

Final sweep is running. Before the results, the scoping result matters more than I expected, because it permanently removes three of your eight items.

What the deposit's read structure forecloses — verified, not assumed

I checked the reads directly rather than assuming:

• Zero soft-clipped reads — 0.0000% of 132,965 reads on chr20. bwa aln aligns end-to-end, so soft clips don't exist in these files.
• Zero supplementary alignments — no split reads anywhere.
• 0.06% of reads flagged paired (106,389 of 167 M), so no usable discordant-pair signal.
• Non-pure-M CIGARs are 0.88%, all small indels.

Structural variants are normally detected by three signals: split reads, discordant pairs, and read depth. Two of the three do not exist in this deposit. So your "structural hotspots" item — soft-clipping, discordant pairing, mapping-quality structure around breakpoints — cannot be done at all, and neither can inversions or any balanced rearrangement, since those are copy-neutral and only visible through breakpoint reads.

Depth is the only structural signal available. That's a second permanent ceiling alongside the mapped-reads-only one you already identified.

Honest scoping of your eight items

Running now

Genome-wide depth binning at 50 kb across all five BAMs in parallel. Zana and Khwit normalised against the three-Viking panel for technical normalisation only — not as ancestry controls, per your instruction. Shared Zana→Khwit events will be flagged separately from sample-specific ones, since the mother–son pair gives inheritance as a built-in filter, which is exactly the thing that makes a shared rare event worth more than a one-sample blip.

Candidate rules fixed in advance, from the lessons already paid for: minimum read support per window, no single-window calls, local coverage and mapping-quality inspection before anything is believed, and ≥3 same-sex controls on X — the SOX3 false positive came from having only two.

The highest-value category remains the one you named: large shared structural variants, rare in modern humans, inherited Zana→Khwit. That's precisely what an ancestry-focused analysis would miss.

I'll report when the binning completes and the analysis runs. If it comes back boring, I agree that closes Zana/Khwit and the investigation is ready for your publication decision.

Zero false positives in leave-one-out, 36 shared windows, and every single one is a loss. That asymmetry is itself the finding to interrogate — let me look at the raw counts rather than the ratios.