@thebusiness7 rotating models have a union of competing sources and fixed references as outgtoups.
English
Insights on Ancestry
41 posts
Insights on Ancestry
@IOAncestry
We help people who've tested their DNA learn about their ancestors from the past.
加入时间 Eylül 2025
4 关注55 粉丝
Modeling a Palestinian Arab requires heavy input from local Levantine sources like Israel_MLBA, Lebanon_IA and Jordan_LBA. Check out the rotation modeling results.
They show a mixture of Levantine, South Caucasian and African ancestry. 🧬
English
@thebusiness7 ... Russia_Karelia_Mesolithic,Iran_GanjDareh_N,Turkey_N,China_Xinjiang_Xiaohe_BA,ONG,Jordan_PPNB,Mongolia_N
Indonesia
@thebusiness7 Mbuti,Lebanon_MBA,Iran_DinkhaTepe_BA_IA1-1,Iran_DinkhaTepe_BA_IA2-2,Iran_HajjiFiruz_BA,Israel_MLBA,Armenia_NerkinGetashen_MBA,Israel_C,Syria_EMBA,Syria_Umayyad,Armenia_KarmirBlur_IA_Urartian,Russia_Samara_EBA_Yamnaya,Serbia_IronGates_Mesolithic,Georgia_KotiasKlde_Mesolithic, ...
Indonesia
@ziad_chjrawi Please provide a list of outgroups and sources (not moderns) you think are better. We’ll re-run the script to find the best models. The current model is already consistent with the distal one published in academia by the way.
English
@IOAncestry Your model simply doesn't account for any of the historically verified demographic changes during the last 2000 years (many of which are quite recent). Your outgroups don't look particularly strict either.
English
There's no better way to rank admixture models via qpAdm than using a rotating strategy. The results are 100% reproducible as we share every parameter including the list of samples used in the analysis with the customer.
Such a stringent approach eliminates all bias/agenda.
Insights on Ancestry@IOAncestry
@ziad_chjrawi A model tested under a rotating approach is more credible than any other base/static reference model. Check Table S32. Of the 38 passing models in their simulated exercise, top 32 p-values are for true model, bottom 6 p-values pass bad models.
English
@ziad_chjrawi A model tested under a rotating approach is more credible than any other base/static reference model.
Check Table S32. Of the 38 passing models in their simulated exercise, top 32 p-values are for true model, bottom 6 p-values pass bad models.
English
@ziad_chjrawi Using a rotating strategy is the most robust way of ranking various admixture models in an unbiased way. The algorithm picks the best model.
Read academic.oup.com/genetics/artic… for more info
English
Our recent post about Palestinians provides a deeper insight into the Levantine genetic profile. It validates the older distal model used in published population genetics papers.
There’s no debate, our model is optimal.
Insights on Ancestry@IOAncestry
@DangitShaun Our model provides a higher resolution picture. EHG in Levantines is from Yamnaya > Armenia as our models showed. The African ancestry is from ETH_Mota4500BP > TZN_PN. The rest of it is local Levantine. paper link: pmc.ncbi.nlm.nih.gov/articles/PMC84… cc @DangitShaun
English
@DangitShaun Our model provides a higher resolution picture. EHG in Levantines is from Yamnaya > Armenia as our models showed. The African ancestry is from ETH_Mota4500BP > TZN_PN. The rest of it is local Levantine.
paper link: pmc.ncbi.nlm.nih.gov/articles/PMC84…
cc @DangitShaun
English
@DangitShaun it’s a proximal model so it doesn’t need pure distal sources. papers usually use the Ethiopia_4500BP to model middle easterners by the way so using a newer admixed source is fine in proximal models.
English
We are making G25 coordinates of about 2300 individual samples of Indians and other subcontinentals from academic studies completely public. The majority of them were on the ~584k SNPs HO panel, they will work nicely with other coordinates.
Link: drive.google.com/drive/folders/…
English
These coordinates are being made public for the first time ever. Amateurs can use them for their PCA and admixture analyses on Vahaduo.
All of these samples are from Nakatsuka 2017, Reich 2009, Moorjani 2013 & Kumar 2023. We hope they help in making Indian popgen interesting.
English
They work perfectly with other publicly available G25 coordinates.
These are not to be confused with the previously publicized low quality Nakatsuka 2017 samples in a recent study.
English
Insights on Ancestry 已转推
Distances with the coordinates I generated. This is pretty accurate! The only thing G25 is best at, is admixture %s. Everything else, like PCA and distances can be imitated.
Since IllustrativeDNA isn't using G25 anymore, I don't understand the point of paying for it.
Anveṣaṇam@vicayana
made my own PCA based admixture calculator using all the HO samples, not bad to be really honest. just needs some fine-tuning and one can have their own calc. imagine charging €26.99 for this lol.
English
@kvali_app That qpAdm model for a Palestinian person was ran by us 🙌
English
That qpAdm model for a Palestinian person was ran by us 🙌
Kvali@kvali_app
qpAdm and G25 give basically the same Palestinian model when fed identical sources. One is 'academic', while the other gets called 'pseudoscience' by pretentious pseudointellectuals. The actual numerical difference? Tiny.
English