Tag Archives: dendrogram - Page 2

Ref1 South Asians + Harappa PCA Clusters

Using the PCA results of the South Asians in Reference I as well as Harappa participants, I ran a couple of clustering algorithms.

First, I scaled the principal components by the respective eigenvalues.

Using Euclidean distance for hierarchical clustering with complete linkage, here's the dendrogram for the Harappa Project participants.

You can compare this to the Admixture-based dendrogram:

The most obvious thing is that I (HRP0001) am an outlier by far.

We inferred three major clusters with the admixture results. Those are intact, though changed a little.

I also ran MClust on the PCA data. The optimum number of clusters was 14. The resulting cluster assignments can be seen in a spreadsheet.

For the Harappa Project participants, the numbers give the probability of assignment to a cluster. For example, for HRP0009 there is a 72% of belonging to cluster 4. For the reference populations, the numbers give the expected number of samples assigned to a cluster.

Harappa and Reference I Dendrograms

Looking at the Harappa dendrogram and the dendrogram for reference I, I thought I would combine them to see where our project participants fit.

Then I got more curious. I wanted to see a similarity tree of all the samples in reference I (2,654) plus the 40 Harappa participants I have processed till now. That came out to be such a huge tree it was impossible to save it in a way to be legible. Finally I compromised by selecting only the South Asian samples from the Reference I dataset and putting them together with the Harappa data. Unfortunately, that doesn't give the Iranian and European-admixed participants any information. I'll have to analyze those separately.

Anyway, here's the South Asian Admixture Dendrogram in PDF format. That means you can search for "HRP" to find all the project members, which is why I like PDF in this case better than an image.

Note that Singapore Indians are such a good stand-in for South Indians.

Harappa Admixture Dendrogram

Using the ancestral component percentages from the Admixture run at K=12 for Harappa Project participants, we can calculate the pairwise Euclidean distance between them. These distances can be used to create complete linkage (i.e. furthest neighbor) hierarchical clustering, which you see below.

Note that this is not a phylogeny. It just visualizes the closeness of your admixture results to others.

Thus in terms of admixture results, the Punjabis mostly cluster together along with the Rajasthani (HRP0033), except for my family (HRP0001 and HRP0035) who cluster (not so closely) with the Sindhi-Balochi guy (HRP0039) likely due to the Southwest Asian and African components.

Interestingly, the Bihari Brahmin (HRP0003) is very different from the Bihari Kayastha participant (HRP0032). The Caribbean Indian samples (HRP0027 & HRP0028) cluster with the Bihari Kayastha, so we can't really say for sure where from India their ancestors originated from.

The South Indian Brahmin samples seem to vary consistently from the non-Brahmin ones.

The Iranians cluster closely except for the Khorasanian HRP0034 and Assyrian HRP0010. The Assyrian Iranian sample is actually closer to the Iraqi/Egyptian Jewish sample (HRP0037) than to other Iranians.

The participants with recent European admixture cluster very loosely with each other. Other techniques will need to be used to pinpoint their specific South Asian origins.

If we make a cut at about 0.3 on this tree, we get 3 South Asian clusters:

  • the Northwest of South Asia
  • South Indian Brahmins, Bihari Brahmin, UP Brahmin
  • South Indian non-Brahmin, Bihari non-Brahmin, Bengalis, Caribbean Indians

I wish I had a thousand South Asian samples to play with. I wonder how this dendrogram would look in that case.

Fst for Reference I Admixture K=12

I had posted the Fst divergences between the estimated ancestral populations for the admixture analysis on Reference I dataset. But a picture is worth a thousand words and this dendrogram (using complete linkage) shows the Fst numbers fairly clearly.

Remember this is not a phylogeny.

Reference I Dendrogram

Handschar created a dendrogram using a hierarchical classifier based on K=12 admixture results and wondered:

When I run a classification based on simple euclidean distances (not a phylogeny), the Armenians and Turks, as they were, prior to the removal of the four North European admixed Behar samples in David's runs, cluster together. The North European component, in Dodecad Armenians, is practically nonexistent. I am not sure how the Harappa project "European" component translates to Dodecad components. If the admixed Armenians are included, it is possible their inclusion is impacting the Armenian population component percentages. Then again, even if included, perhaps your runs are picking up on something not previously detected. The Armenians, in previous classification runs, ordinarily matched one or more of the Caucasian Jewish groups.

While looking into his question, I figured that I would create some dendrograms too. The ones here are based on the K=12 admixture results of Reference I dataset (spreadsheet). Also, I am using the pairwise Euclidean distance of the Admixture results between population groups to do a complete linkage hierarchical classification. So these dendrograms show which groups are closest in terms of their admixture percentages and do not show shared ancestry. In other words, it is not a phylogeny or a family tree.

First, I used the mean admixture percentages for each group, as given in the spreadsheet.

Reference 1 Mean Admixture Complete Linkage Dendrogram

There are a number of outliers in the dataset. For example, some Arabs and Sindhis with African admixture, some Armenians with a lot more European component than the rest, etc. Therefore, I thought a better approach would be to do the same classification using the median admixture percentages for each population group.

Reference 1 Median Admixture Complete Linkage Dendrogram

Using the median sample from each population, handschar was correct that the Armenians match the Caucasian Jewish groups.

UPDATE: Here's another dendrogram in which I take the mean of the ancestral components for each population after removing outliers.

Reference 1 Mean (No Outliers) Admixture Complete Linkage Dendrogram

Again, don't take these dendrograms to heart. All they show is the distance between the admixture results of different populations.