Affy to Illumina conversions complete?

[nolnacsj]

Fair enough EXCEPT the FTDNA site says this:

We are pleased to announce that the conversion of all Family Finder results to the Illumina Omni Express platform is now complete. Please note, this means your Affymetrix results will be removed from the system on Tuesday, May 10th, and will no longer appear on your "myFTDNA" page.

I noticed that FTDNA changed the date that the Affy results will disappear, it has now been extended to Tuesday May 10th. It used to say that it was today, May 3. I am glad the Affy's will remain, at least a little longer, but hopefully they will end up remaining longer than that.

Judy

[mkdexter]

Today I compared raw data from an Affy match I lost in Illumina. I find this interesting as (on the surface) I see no reason the match was lost. It must have been a longest block lower than 7.something cM.

FTDNA Affy Info:
5th Cousin - Distant Cousin - 32.37 7.87

Using David Pike's utility:

Affy 13 segments, longest block =
Chr 2 position 86869529 to position 100415547 ( 13.55 Mb)

Illumina, 27 segments, longest block =
Chr 2 position 86866217 to position 100415100 ( 13.55 Mb)

I sent the full report to FTDNA and can't wait to see what they say.

MD.

[mkdexter]

Correction _ found the loss

I checked and noticed that FF Affy shows this match as chromosome 5 as 7.87cM, not chromosome 2 as show quoted below.

The chromosome map data and the Pike utility comparison don't even match. Everything is all different. I'll have to figure out why FF Affy chromosome browser shows different data.

Matt.

Today I compared raw data from an Affy match I lost in Illumina. I find this interesting as (on the surface) I see no reason the match was lost. It must have been a longest block lower than 7.something cM.

FTDNA Affy Info:
5th Cousin - Distant Cousin - 32.37 7.87

Using David Pike's utility:

Affy 13 segments, longest block =
Chr 2 position 86869529 to position 100415547 ( 13.55 Mb)

Illumina, 27 segments, longest block =
Chr 2 position 86866217 to position 100415100 ( 13.55 Mb)

I sent the full report to FTDNA and can't wait to see what they say.

MD.

[gtc]

"your lost matches was"

Thanks for the info Matt.

I'd have expected better grammar from FTDNA.

[mkdexter]

I checked and noticed that FF Affy shows this match as chromosome 5 as 7.87cM, not chromosome 2 as show quoted below.

The chromosome map data and the Pike utility comparison don't even match. Everything is all different. I'll have to figure out why FF Affy chromosome browser shows different data.

Matt.

Per my research this one Affy match I had was not a match. I tried and tried to lower the tolerances in the Pike utility and no where did I see a match of 589 SNPs in the Affy data at the position FF reported on chromosome 5.

In this one case either the chromosome map is reporting wrong, or the Affy test said there was a match and there was not a match whereas the Illumina test reported no match from day one.

Matt.

[goshelli]

Lost Affy matches with 10 cM + shared segments

Two of my 4th to distant Affy matches that were not picked up by Illumina each had longest shared segments over 10 cM.

#1 30.91 shared cM, 10.15 cM longest shared segment, 7 shared segments
#2 27.18 shared cM, 13.54 cM longest shared segment, 7 shared segments

Has anyone else lost matches in the 10 cM + range? Isn't 10 cM+ over the threshold for IBD vs IBS? I'm looking for an educational assist if anyone has an explanation.

[Ann Turner]

I'd have expected better grammar from FTDNA.

Re "Your lost matches was" -- I bet I know exactly how that happens, because I often do it. I have canned paragraphs that I need to edit in several places depending on whether a singular or plural is required, and I don't always catch all the words that need changing. I aggravate myself no end when I see what I've sent out

[Ann Turner]

I checked and noticed that FF Affy shows this match as chromosome 5 as 7.87cM, not chromosome 2 as show quoted below.

The chromosome map data and the Pike utility comparison don't even match. Everything is all different. I'll have to figure out why FF Affy chromosome browser shows different data.

Matt.

This wouldn't apply to your investigation because the segment is on an entirely different chromosome, but I just thought I'd mention that the chromosomal position of a match can easily vary between Affy and Illumina. The SNP positions can be different because Affy uses build 36 of the human genome and Illumina uses build 37. I've seen cases where that doesn't seem to be an adequate explanation, though.

[Javelin]

Re "Your lost matches was" -- I bet I know exactly how that happens, because I often do it. I have canned paragraphs that I need to edit in several places depending on whether a singular or plural is required, and I don't always catch all the words that need changing. I aggravate myself no end when I see what I've sent out

I use highlighting for that in MS Word. It's not foolproof, but it does help me remember to look for those.

[mkdexter]

This wouldn't apply to your investigation because the segment is on an entirely different chromosome, but I just thought I'd mention that the chromosomal position of a match can easily vary between Affy and Illumina. The SNP positions can be different because Affy uses build 36 of the human genome and Illumina uses build 37. I've seen cases where that doesn't seem to be an adequate explanation, though.

Yes. The fact is something is wrong. The raw data and the chromosome map readings do not match for Affy raw data. In the FF Map the longest cM is on chromosome 5 - and in David Pike's comparison its on chromosome 2 for both Affy and Illumina in the same general area, not exact but close.

Either FTDNA did not provide the raw data (they strip some of the medical SNPs out that are used in matching) or there is a problem with Affy FF matching and it just showed up while I was looking at it.

What bugs me is that I can't even validate an Affy match.

I've heard nothing for a day from FTDNA.. today will be interesting.

[Ann Turner]

I use highlighting for that in MS Word. It's not foolproof, but it does help me remember to look for those.

Good idea, although I use a "keyboard extender" called ShortKeys. I type a couple of characters and ShortKeys pastes in long paragraphs or passwords or whatever. I couldn't live without it.

[JohnG]

Yes, I did lose my confirmed third cousin. We know we are third cousins, and had recently determined that on paper before being matched on Affy. He was my closest match overall. All those SNPs must have been in the Affy spots and not in the Illumina.

Your case would seem to be an excellent one to report. There should be no question with third cousins. Have you reported it to FTDNA with kit numbers?

Also if your third cousin will send you their raw data you could do a comparison of the Illumina results at David Pike's site. Ideally you would compare your Affy results and then your Illumina results. That site is

http://www.math.mun.ca/~dapike/FF23utils

[Javelin]

Good idea, although I use a "keyboard extender" called ShortKeys. I type a couple of characters and ShortKeys pastes in long paragraphs or passwords or whatever. I couldn't live without it.

That's a great idea. I usually just have my canned responses in Word docs, then alter the highlighted text, but it does take a little work to open the docs then cut, paste, and alter.

[Ann Turner]

Yes, I did lose my confirmed third cousin. We know we are third cousins, and had recently determined that on paper before being matched on Affy. He was my closest match overall. All those SNPs must have been in the Affy spots and not in the Illumina.

Third cousins typically have several segments > 5 cM in length (usually longer, in fact). Did that hold true for you? Even if one segment doesn't make it through, other segments should still have a chance. Have you checked that your cousin has actually been converted?

[mkdexter]

Swimmingly I sent you a private message note.. Please check it and reply if you are ok with answering.