As I have mentioned before, cousin marriages were fairly common among my family. My parents are first cousins. So are my father’s parents. My mother’s parents are second cousins once removed. So instead of 32 great-great-great grandparents, I have only about 18.
Since my wife and I are not related, I wondered how my inbred genome had transmitted to our daughter.
Using David Pike’s ROH utility, I computed the regions of homozygosity for my parents, me, my wife, and my daughter, all tested by 23andme.
I used the default settings for the utility. The total Mb gives the total size in megabases of the long autosomal regions where both alleles are the same. The longest ROH gives the size of the longest such region. Percent Homozygous is the percentage of the genome where the two alleles are the same.
I included the worst chromosome column because of my chromosome 9, which is beyond crazy. This column gives the percent homozygosity of the worst chromosome.
||Longest ROH (Mb)
||Worst chromosome (%)
As you can see, my Dad has higher levels of homozygosity than my Mom as expected and I have the highest levels. My wife is not inbred at all and our daughter has ROH results about the same as my wife. So one generation of marrying someone unrelated, even if from the same/similar ethnicity, has removed all the long runs of homozygosity bred over generations. Good news!
A few months ago, I made my DNA genotyping results from 23andme public.
Since I got results for both my parents as well, I have now used BEAGLE to phase my genetic data. In simple words, I have been able to separate the contribution of my Dad and my Mom on my DNA.
I am making my phased genome public too. It’s in Plink format.
I haven’t made much use of the phased genome yet. So if you have any ideas about what can be done with a phased genome, please let me know.
I have also pledged to make my full sequenced genome public when genome sequencing becomes cheaper and I get it done.
Four years ago, I built my desktop computer. Now it was getting a bit long in the tooth, so I decided to upgrade some of its parts.
- Motherboard: From Abit IP35E to Gigabyte GA-Z68XP-UD3
- Processor: From Intel Core 2 Duo E6420 to Intel Core i7-2600
- Memory: From 2x1GB PC2-6400 DDR2 to 2x4GB PC3-14900 DDR3
- DVD Drive: From PATA to a SATA DVD Writer
The computer’s now fast and powerful. My Harappa Ancestry Project analyses run much faster than before.
Ubuntu had no trouble running after the change. However, Windows XP refused to boot and I have to do a reinstall.
I posted my genetic ancestry results. Now, we’ve got my parents, my sister and my wife tested with 23andme. So I thought a comparison would be interesting.
Here’s the ancestry painting from 23andme which uses three reference populations: Yoruba from Nigeria, Chinese and Japanese, and Utahns of Northwestern European descent.
You can basically use my wife as a sort of reference for Punjabi ancestry here (which is 3/4th of our ancestry too). Also, my wife and I are unrelated.
As you can see, while our results are close, my mom and sister have more African and I have the least.
And here are the similarity numbers for us with different reference populations.
|Central & South Asians
As compared to my wife, we are closer to Africans and farther from Eastern Asians, Native Americans (who are really a branch of East Asians) and Oceanians.That’s expected because of the 25% Egyptian ancestry we have.
Finally, here are our Dodecad Project results.
Similar results but interesting differences.
I tested with 23andme in April 2010 and then upgraded to their version 3 chip with almost a million SNPs last Christmas.
Now I am releasing my personal genome in the public domain.
To the extent possible under law, Zack Ajmal has waived all copyright and related or neighboring rights to Zack Ajmal 23andme v3 Genome. This work is published from: United States.
You can download my genome data in zipped files:
Razib has a list of people who have made their 23andme genomes public.
When Blaine Bettinger released his genome into the public domain, he issued a challenge:
So, I’m challenging everyone who reads this to download my data and analyze it to find the most interesting or surprising results. For example, you could use my most recent 23andMe V3 data.
I’ve already done a fair amount of analysis myself, including the Promethease reports above (and see here), and a recent blog post about my vastly increased Type 2 Diabetes risk. However, perhaps there’s a recent but relatively study that applies, or perhaps there’s a story you can weave with a handful of SNPs. Or, even better, what can you tell me about my ancestry other than mtDNA and Y-DNA haplogroups? Don’t worry about the strength of the study, reproducibility, etc. – I’m aware of the uncertainties associated with this type of research, and my goal here is to make people aware of possibilities.
Please post your findings in the comments below, and in two weeks I’ll pick the most surprising or interesting findings and make them the focus of a new blog post.
Can you surprise me with my own genome?
I have done a fair amount of analysis on my genome. For example, here’s my Promethease report. My ID is DOD128 in Dodecad, PKEG1 in Eurogenes and HRP0001 in Harappa.
My challenge for you would be to find interesting information about my chromosome 9 which is 93% homozygous.
If you analyze my genome, it would be great if you could let me know about what you found as I am always hungry for more information.
Dodecad has come up with a new version (v3) of its admixture results. Here are my results:
Dodecad also has a fun tool to check one’s results against different population averages. My closest populations are:
||Bene Israel Jews
If I make use of mixed mode, the tool tries to find a combination of two ethnic groups with differing percentages that fits my results best.
||Two Population Mix
||17.3% Palestinian + 82.7% Sindhi
||17% Morocco Jews + 83% Sindhi
||17.3% Palestinian + 82.7% Punjabi Arain
||17.2% Egypt + 82.8% Punjabi Arain
||82.9% Sindhi + 17.1% Egypt
||17% Lebanese + 83% Sindhi
||16.7% Jordanians + 83.3% Sindhi
||16.7% Jordanians + 83.3% Punjabi Arain
||15.8% Samaritians + 84.2% Sindhi
||16.9% Ashkenazi + 83.1% Sindhi
This actually fits reasonably well with my actual ancestry (75% Punjabi + 25% Egyptian).
When the doctor told me I had Ureterolithiasis, I logged into 23andme to check my genetic risk. There was only one SNP (rs4293393) listed there. The G allele increased the risk 14% but I have AA, so typical odds.
Next step was checking SNPedia where I found 8 SNPs, of which some are given below.
rs219780 (23andme): The high risk is CC but I wasn’t genotyped at this location. Also, CC is the most common, so it is quite likely that I have that.
rs219778 (23andme): Carriers of TT have a slightly increased risk and that’s what I have.
rs9310709 (23andme): Risk allele is C and I have CC.
rs10941694: I was not genotyped.
rs13070584: I was not genotyped.
More important than these though is the simple fact that my Dad had it too. Thus if there is a genetic association, I am likely to be higher than typical risk.