Affy to Illumina conversions complete?

[mkdexter]

Dexter, you forget that FTDNA stated in an email that those who chose the conversion would be done in 2-3 weeks with the rest of the database completed in 6-8 weeks.
FTDNA know the size of the database, knew how many kits would need converting.
They are the ones who came up with the 8 weeks.
That was back on the 11 Feb, so we are well over 8 weeks.


I think FTDNA have taken on more than they can handle with sales, 111 upgrades and now the DNA Hertiage merge and we are seeing the consequence in delays in results.

No I didn't forget. I still think they've done pretty good considering I know how complicated it is to go from a sample to a web page match. It takes a ton of work for a little tiny bit of DNA to make it to the web site.

New tests and company adds are not necessarily delaying FF tests. These tests run on different machines, some in Tucson and some in Houston.

I'll agree with the fact that the sale added a workload if you mean to the office help. Not the lab though, not yet. When that comes the conversions will be done basically.

When they said 2 to 3 weeks they didn't anticipate 45% of the entire database signing up for the first conversion. The conversion start was Feb 16. That's just over 10 weeks and that isn't bad considering many new tests came in as well and those were also ran, not just the conversions.

Sure everybody can improve. Sometimes the improvement needs to come in the form of understanding what's involved - this stuff is rocket science and its not an overnight process.

[mkdexter]

Kasandra I see we agree on a few things and disagree on some others.

However when you say the chromosome browser is "BS" that's not disagreeing with me, that's arguing with reality.

All our segments do map out either on a paternal ancestry map or a maternal one. If you don't know which segment is which find a way to figure it out.

Certainly someone like yourself with some smarts can come up with a few methods that would help understand what to do if the lineage is not known. You can use other data, trees, match lists, GEDMatch, HIR and known relative comparisons for example.

[mkdexter]

In a nutshell here's the lab process:

1. DNA Extraction/Amplification
2. Fragment/Precipitate/Re-suspend/Overnight Hybridization
3. X-Stain/Scan
4. Data Analysis

After that there's the IT process of loading the data into the computers and letting the algorithm parse the data and find matches.

It all takes about two weeks and that's only if the sample was a good one and didn't have to be re-ran in the lab or re-swabbed by the tester.

MD.

[Kasandra]

Kasandra I see we agree on a few things and disagree on some others.

However when you say the chromosome browser is "BS" that's not disagreeing with me, that's arguing with reality.

All our segments do map out either on a paternal ancestry map or a maternal one. If you don't know which segment is which find a way to figure it out.

Certainly someone like yourself with some smarts can come up with a few methods that would help understand what to do if the lineage is not known. You can use other data, trees, match lists, GEDMatch, HIR and known relative comparisons for example.

I have 'figured it out' the only way these segment matches can be made meaningful (without other close known relatives tested) is if FTDNA (or 23andMe) does it with the raw data. Otherwise, it's a total waste of time. Just because direct line families inherit the same bits of DNA, the same is not true for indirect lines ie cousins. I did a big mathy thing at one point that proved that as long as there are two children or more per generation, it's statistically meaningless to try to track segments through cousins. No matter how smart anyone is, the reality is unless exact segment boundaries are reported, which isn't going to happen, the only folks that can make a realistic determination on which side of the family those segments fall on is the testing companies with access to the raw DNA on both folks.

Please, do not advocate GEDmatch to me as a method of doing anything. I have zero faith in their DNA to DNA comparison system. The only thing that's of solid value is triangulation and that's with the caveat that those folks can be in-laws.

I do use HIR search but even then, it requires others to be willing to upload their data to a third party site. I find a lot of older folks unwilling to do that and with such low response recently, using 3rd party sites is unlikely in any meaningful form. FTDNA needs to be the one to compare raw data and say, these are from your A side these are from your B side, these seem to match both sides and this pile over here, don't have any overlap with each other. It's not that hard it just takes time and resources that they'd rather spend on the Ydna Heritage type things. Just because I have other ways of figuring things out doesn't mean that chromosome segment overlaps are meaningful.

[vivianruth]

Thanks for arguing

Thank you, Matt, Kasandra, rivergirl, and others, who are arguing these things out. I appreciate the way you generally stick to issues on which you disagree (rather than ad hominem arguments, which seem to occur too often on the web) and try to make your points more clearly and vigorously. Sometimes I get lost. But, generally, I learn more an more about this stuff. Your disagreements on this forum provide real meat to chew on.

vivian

[nathanm]

I replied inside the quotes. The only real plus from my point of view is the ability to sort. And 'maybe' YDNA matching for males. Although, you can match the YDNA of your RF and AF matches just by downloading the spreadsheet and sorting it to your hearts content. I really don't think having the MTDNA matching across the service comes close to having 500 YDNA and FTDNA haplogroups to play with. But that's just me.

All you can compare with 23andMe data is the haplogroup and sub-clade. But two people in the same sub-clade could have vastly different haplotypes. Besides, there are lots of potential Y-DNA matches who wouldn't ever find each other, because they share no detectable blocks of autosomal DNA.

[nathanm]

>Here I don't agree with you. I've used MTDNA and Y DNA haplogroups to good effect on 23andMe to identify which cousins are from what side of my genome. I can pretty much predict 40% of my matches with greater than 90% accuracy based on haplogroup alone. This allows me to tailor my invitations but even better I've been able to analyze what percent of my matches are from which grandparent. The overall numbers are slightly less skewed than my reply numbers because I have one whole section of my genome made up of polite Canadians .. they seem more likely to respond than most other groups.

You mother's and father's sides of the family must be from very different places then, because that's only one of two ways the haplogroups would be useful for the other 30 family lines. (The other is testing several members of your family from different lines, but that gets expensive, and some people are still reluctant to submit to a DNA test.) If you have typical European haplogroups like most customers both here and at 23andMe, matching haplogroups can give you false-positive leads, and non-matching haplogroups can keep you from investigating false-negative leads.

And, it's not as big a change as you make out. I've run DNA tests for years. Especially since I don't believe they've tweaked the algorithms at all, that's why so many cousins are just clumped in the 'I dunno' category of something to distant.

Yes, I did a PCR lab in my college biology class too. But the difference between platforms isn't in the testing, which is mostly automated anyways. The big change is in how much data is collected, the orientation of that data (both companies are using the same platform now, but sometimes reporting the data differently), and figuring out which SNPs to compare. The algorithms they had to update, which I've mentioned elsewhere, were for PF. I don't know if they had to make any changes for FF. As for the 4th or 5th to distant cousin category, they're just giving an estimate. We probably have more FF matches in their system, that just don't meet the threshold. I don't think 23andMe is any different on that front.

> OK, I just said I wasn't impressed by it. FTDNA evidently was impressed by it ergo the big press release.

That's standard practice for any business. There might even be some regulatory requirement to announce any mergers or acquisitions, but I'm just guessing. I know it is required for publicly traded corporations.

[mkdexter]

Thank you, Matt, Kasandra, rivergirl, and others, who are arguing these things out. I appreciate the way you generally stick to issues on which you disagree (rather than ad hominem arguments, which seem to occur too often on the web) and try to make your points more clearly and vigorously. Sometimes I get lost. But, generally, I learn more an more about this stuff. Your disagreements on this forum provide real meat to chew on.

vivian

Is that it? Time to stop then. I can't do anything more to help the person who just wants to argue and not learn. Like I already said its not disagreeing with me, its disagreeing with fact (chromosome mapping).

[mkdexter]

There was a comment about FTNDA had not tweaked the algorithms in FF. I'm not going back and find it to put it in quotes.. it wasn't fact - it was opinion. It was opinion given without the facts actually.

The fact is for part of a day we had two false 2nd cousins because their sums were so high their tests fooled the algorithm. FTDNA found and fixed it that same day. So yes when they need to they do adjust the FF algorithm.

They also plan on adjusting it again. I provided FTDNA some new segment persistence data and they told me they'd be adjusting things soon.

That's facts not opinions.

Another fact. The" I dunno" category has very little to do with the algorithm. It has to do with segment persistence issues. This is the very data I provided to FTDNA recently. Again a statement made about how its the algorithm's fault without all facts.

I'll be providing a lot more data on this stuff like persistence and how its influencing FF results soon. I have in fact already mentioned some of this on the forum in other posts. Its all there and free.

Anyway, those are the facts.

MD.

[mkdexter]

Affy conversions to be completed this week

I have a note from FTDNA that says the Affy conversions will be completed by this week. Probably (and we all know how things change) by Wednesday and uploaded Thursday. That's by April 28th.

FF Affy results will be removed as far as I know. Save them if you have not already and want to keep them. As far as I'm concerned they are still matches even if they don't show up again. Of course going forward they won't be used for any comparing in Family Finder so its not necessary to keep them unless you want to for archival purposes.

[CAM]

Affy to Illumina

I have a note from FTDNA that says the Affy conversions will be completed by this week. Probably (and we all know how things change) by Wednesday and uploaded Thursday. That's by April 28th.

FF Affy results will be removed as far as I know. Save them if you have not already and want to keep them. As far as I'm concerned they are still matches even if they don't show up again. Of course going forward they won't be used for any comparing in Family Finder so its not necessary to keep them unless you want to for archival purposes.

Matt,
Thank you for the very important warning about keeping Affy results for those of us who plan on doing so.
Cheryl

[nolnacsj]

I have a note from FTDNA that says the Affy conversions will be completed by this week. Probably (and we all know how things change) by Wednesday and uploaded Thursday. That's by April 28th.

FF Affy results will be removed as far as I know. Save them if you have not already and want to keep them. As far as I'm concerned they are still matches even if they don't show up again. Of course going forward they won't be used for any comparing in Family Finder so its not necessary to keep them unless you want to for archival purposes.

Matt, please tell me that there will be a big final giant dump of Affy matches into Illumina when the conversions are complete, as only 7 of my 33 Affy matches have shown up in Illumina so far! I did not expect to loose 26 of my original matches I hope this is not the final tally......

Judy

[mkdexter]

Yes this will be the final dump of converted kits into Illumina. I don't know how many are in it.

[gtc]

only 7 of my 33 Affy matches have shown up in Illumina so far!

That's very low.

I have 28 of 52. It's hard to believe that the remaining 24 will suddenly pop up in the next 2 days or so.

[Javelin]

I have a note from FTDNA that says the Affy conversions will be completed by this week. Probably (and we all know how things change) by Wednesday and uploaded Thursday. That's by April 28th.

I'll believe it when I see it -- I'm still waiting for my Batch 400 results, which were supposedly due to have been processed by the end of February when I signed up for the conversion.