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

Yes — this is the first genuinely interesting control result, and Outcome A clearly happened. The unrelated Viking pairs do not behave like Zana–Khwit: at depth 8+, ZK is down to 1 OH per 223,119 sites, while the unrelated pairs sit around 1 per 12k–21k, and the depth-2 values were nearly identical across all four comparisons. That is exactly the qualitative pattern we wanted. fileciteturn7file0L104-L124

I would not quite let Claude close Phase 2 with the “siblings excluded at p ≈ 6×10⁻⁶” sentence yet, though. That p-value is more model-dependent than the rest of the result. He estimated a sibling expectation as roughly one quarter of the unrelated OH rate, but our unrelated controls are cross-population and share individuals across comparisons. Population differentiation raises the unrelated OH floor, so using that floor to derive a sibling expectation can make the sibling test look stronger than it really is. Claude already acknowledges the cross-population caveat, but the p-value headline outruns that caveat. fileciteturn7file0L126-L147

So I’d say the correct conclusion right now is:

Our custom depth-stratified OH test successfully distinguished Zana–Khwit from three documented-unrelated ancient-human comparisons under the same analysis pipeline. The Zana–Khwit pattern is strongly consistent with parent–offspring and agrees with the published R0 result by an independent route.

That is already a big win. No need to squeeze an overprecise significance claim out of it.

And this is exactly why we identified the harder same-site control earlier. Since Outcome A happened, I would now run one matched same-population test before unlocking Phase 3. Best choice remains the Hofstaðir trio:

• VK110 + VK230 = documented first-degree
• VK102 + VK230 = no documented relationship
• same site, same period, same pipeline
• shared VK230 controls a lot of nuisance variation

If that matched negative retains an OH floor while the known first-degree pair collapses toward zero, then I’d consider the relatedness calibration properly nailed down.

So my recommendation to Claude would be:

Do write the Outcome A result into the README/DISCARDED record, but phrase sibling exclusion as provisional and do not open Phase 3 yet. Run the Hofstaðir matched control next. If it reproduces the relationship-dependent behavior, close Phase 2 and open Phase 3.

Also, the download/repair side is now legitimately done: all three BAMs are ~8.5–9.4 GiB, MD5-matched, full-decoded, read counts exactly match the archive, and all are indexed. fileciteturn7file0L68-L94

So we’re very close. The core null test passed. One tougher matched control is the clean way to finish this.

can we get back to bigfoot, viking dna is interesting but 30gb downloads and 10 hours of processing is not fun, we are running out of time and tokens

Yes. I think we have enough Viking DNA for the purpose it was serving. The three-control experiment gave Outcome A: our method can distinguish Zana–Khwit from unrelated ancient humans. I would freeze that branch now and not spend another 7–10 GB and hours on Hofstaðir unless some later Bigfoot result actually depends on tightening that calibration.

We can finally get back to the thing we were trying to investigate in the first place. The highest-value next move is the Ketchum material, because it’s already sitting on disk and is tiny compared with these BAMs:

1. Samples 26, 31, 140 — independently classify the surviving nuclear consensus sequences against human, black bear, and dog. Hart reportedly got roughly bear / human / dog; we should reproduce or falsify that ourselves and quantify how decisive each assignment actually is.
2. For anything that looks human, ask the question that matters for H4: is it merely ordinary human sequence, or is there anything incompatible with ordinary Homo sapiens? The consensus-only limitation means we cannot recover mixtures, contamination, heterozygosity, or coverage, but we can still test what the sequences themselves are.
3. Then return to the Sykes Texas sample 25072, the one that returned human mtDNA. Pin down everything recoverable about its provenance, HVR1 result, sequence/accession availability, and whether any nuclear material/data survive.
4. Build a short table of the best alleged Bigfoot biological samples that returned human rather than bear/deer/dog/etc. Those are the cases H4 actually predicts are interesting. Ignore piles of obvious animal results unless they serve as controls.

That gets us back to the central question:

Are “human” Bigfoot DNA results simply contamination/misidentification, or has anyone ever produced a sufficiently deep human-looking sample that could reveal an unusual human lineage?

And the Ketchum consensuses are a nice immediate test because we can probably answer the first part without another damn 30 GB download. 😄

I’d tell Claude simply:

Freeze Viking controls. Outcome A is sufficient for current calibration; no Hofstaðir download. Return immediately to Bigfoot. Next task is independent species classification of Ketchum consensus samples 26, 31, and 140 using the existing local archive, with Qwen 5 analysis and Qwen 7 red-team review. No large downloads. Report whether Hart’s bear/human/dog assignments reproduce, how decisive each assignment is, and what the consensus-only data can and cannot establish.

That’s where I’d go next.

so where are we?

We’re basically at the end of the calibration detour and ready to get back to actual Bigfoot DNA.

The Zana/Khwit branch did its job. We independently reproduced the published basics: coverage, ancient-DNA damage, sex, mtDNA haplogroup L2b1b, and a relatedness signal that behaves like parent–offspring. Then the three unrelated Viking controls gave Outcome A: unlike Zana–Khwit, they retained a clear opposite-homozygote floor at higher depth. That means our custom relatedness test is at least useful enough for calibration.

I would now freeze the Viking branch. No more 30 GB downloads unless a later result truly depends on them.

The next Bigfoot-specific branch should be the Ketchum material already on disk. We have three surviving nuclear consensus sequences:

• sample 26 ~2.73 Mb
• sample 31 ~0.53 Mb
• sample 140 ~2.10 Mb

The next task is to classify those independently against human / black bear / dog and see whether Hart’s reported assignments reproduce. That is the fastest route back to the central question:

Are these alleged Bigfoot sequences actually unusual human-lineage material, or do they reduce to ordinary known species?

Important limitation: because these are consensus sequences and not raw reads, we cannot reconstruct contamination, mixtures, coverage, allele balance, or heterozygosity. But we can still ask what species the surviving sequence itself best matches.

So where we are now is:

Calibration branch: effectively complete.
Bigfoot investigation: resume with Ketchum 26/31/140.
Phase 3 anomaly hunting on Zana: not worth prioritizing right now.
Bigfoot remains unpublished.

If you want, I can give you the exact short handoff to Claude for the Ketchum classification next.

what does a clear opposite homozygote floor at higher depth mean?

It means this:

At any given DNA position, each person usually has two copies of a base, one from each parent. If one person is AA and the other is GG, that is an opposite homozygote: both are homozygous, but for different alleles.

For a true parent and child, that should basically never happen at a real autosomal site, because the child must inherit one of the parent’s two alleles. If the parent is AA, the child has to carry at least one A.

At low sequencing depth, though, the data are noisy. You can get fake opposite homozygotes because one allele was missed, a read was wrong, ancient-DNA damage occurred, or the mapping was imperfect.

As depth gets higher, those fake OH calls should disappear.

That is what happened with Zana–Khwit:

• depth 2: about 1 OH in 2,332 sites
• depth 8+: about 1 in 223,119

So their OH rate kept collapsing toward zero.

The unrelated Viking pairs behaved differently. Their OH rate also improved with depth because some errors disappeared, but then it stopped falling so fast and settled into a nonzero range:

roughly 1 OH in 12,000–21,000 sites at depth 8+.

That stable nonzero level is what we mean by the “floor.”

It says:

“Even after sequencing noise is mostly removed, these two people still genuinely have positions where one is AA and the other is GG.”

That is normal for unrelated humans.

So visually, think of it like this:

Parent-child candidate lots of apparent differences → fewer → fewer → almost none

Unrelated pair lots → fewer → fewer → then levels off because real genetic differences remain

The Viking “floor” is that leveling-off point.

And the reason it mattered so much is that it showed our pipeline was not simply forcing every pair toward zero at high depth. The unrelated controls retained real opposite-homozygote differences, while Zana–Khwit mostly did not.

so that just proves that khwit was her son

Pretty much, yes. 😄

The Viking detour was not telling us anything exotic about Zana. It was validating that our independent method really can distinguish parent–offspring from unrelated humans.

Once that is combined with the other pieces:

• Zana is female.
• Khwit is male.
• Their autosomes behave as parent–offspring.
• Khwit’s single X chromosome is compatible with Zana; a male gets his X from his mother.
• They carry essentially the same maternal mitochondrial lineage, and mtDNA is transmitted by the mother.

That combination makes Zana → mother, Khwit → son the natural genetic relationship.

So yes: the practical conclusion from all that 30 GB of Viking DNA was basically:

Our independent analysis agrees that Khwit was Zana’s son.

The Vikings were there to make sure our homemade relatedness method wasn’t just producing “parent-child” for everybody. They showed it doesn’t.

Which is useful scientifically, but I agree: we have sufficiently established who Khwit’s mom was. Back to Bigfoot. 😄

Zana, you ARE the mother, lol

I guess this may come in handy as we continue as we can find inherited anomalies in how the genes are regulated?

we khwit hairy like his mom? what happened to him?

and speaking of hairy, what about those 2 twins in mexico that were covered in body hair?