Nebula Genomics

[Ibericus]

Well, it's back to $299. I'm really considering it this time. If I knew for sure that I will be able to upload it to Yfull I would buy it right away.

[teepean47]

For what it's worth, I've just been told by a Nebula customer support agent that they use the hs38 analysis set without ALT/HLA contigs or decoys (ftp://ftp.ncbi.nlm.nih.gov/genomes/a...sis_set.fna.gz), in case you were specifically curious about the reference used.

I was just curious what kind of details the header has.

[Ibericus]

I have a $50 off coupon for the WGS with lifetime subscription. 499-50=$449 at current prices. PM if anyone is interested.

[JMcB]

For those interested, this was posted earlier today on YFull’s Facebook Page.


We received two samples of Nebula Genomics. One from our new client, the other from Nebula team. Here is a comparison with other laboratories. For comparison, samples obtained from March 30, 2020 were taken.

https://scontent-mia3-2.xx.fbcdn.net...55&oe=5F0E02CA

[digital_noise]

For those interested, this was posted earlier today on YFull’s Facebook Page.


We received two samples of Nebula Genomics. One from our new client, the other from Nebula team. Here is a comparison with other laboratories. For comparison, samples obtained from March 30, 2020 were taken.

https://scontent-mia3-2.xx.fbcdn.net...55&oe=5F0E02CA

I’m not very knowledgeable about this, but is it correct to assume higher % the better? If that’s the case, FTDNA is lacking

[JMcB]

I’m not very knowledgeable about this, but is it correct to assume higher % the better? If that’s the case, FTDNA is lacking

While there are others on this thread who a better qualified to answer your question, I would imagine that’s the case.

I will add that in a response to a question on their page - “Width or Depth, which of them (is) more important,” - Vadim Urasin replied: “Width. 15x depth is enough for genealogy, 10x depth is acceptable”.

[Dave-V]

While there are others on this thread who a better qualified to answer your question, I would imagine that’s the case.

I will add that in a response to a question on their page - “Width or Depth, which of them (is) more important,” - Vadim Urasin replied: “Width. 15x depth is enough for genealogy, 10x depth is acceptable”.

I'll be honest I'm not sure what to believe anymore because all of our reported information come from companies who have a vested interest in certain products over others. Even YFull is not strictly neutral anymore.

YFull tends to "call" SNPs at much lower numbers of reads than anyone else so it's not surprising that Vadim would say a 15x depth is sufficient for genealogy. From what I've seen anything under 30x results in large areas of low coverage across regions that would still be counted as "covered" even if the number of reliable SNP calls is lower. And SNP calls are what's important.

You really see that difference on the Y-DNA Warehouse comparative statistics (https://ydna-warehouse.org/statistics.html) when you compare 10x, 15x, 30x, etc WGS callable loci (there defined as 4 or more reads). The callable loci area jumps up quite a bit from lower read depths to higher as the areas with 4 or more reads increases. Big Y-700 there compares favorably to the 30x (or so) WGS testing as far as callable loci goes but I can't quite reconcile that to all the numbers in other vendor reports. Normally I would take the Y-DNA Warehouse as the most reasonably-neutral comparison numbers though. But it would be better if we had some statistics about numbers of SNPs called of different types from the various kinds of testing.

Anyway just based on that I wouldn't go anywhere lower than a 30x myself, but as I said there's a lot of ways of looking at this based on numbers from any one company.

[JMcB]

I'll be honest I'm not sure what to believe anymore because all of our reported information come from companies who have a vested interest in certain products over others. Even YFull is not strictly neutral anymore.

YFull tends to "call" SNPs at much lower numbers of reads than anyone else so it's not surprising that Vadim would say a 15x depth is sufficient for genealogy. From what I've seen anything under 30x results in large areas of low coverage across regions that would still be counted as "covered" even if the number of reliable SNP calls is lower. And SNP calls are what's important.

You really see that difference on the Y-DNA Warehouse comparative statistics (https://ydna-warehouse.org/statistics.html) when you compare 10x, 15x, 30x, etc WGS callable loci (there defined as 4 or more reads). The callable loci area jumps up quite a bit from lower read depths to higher as the areas with 4 or more reads increases. Big Y-700 there compares favorably to the 30x (or so) WGS testing as far as callable loci goes but I can't quite reconcile that to all the numbers in other vendor reports. Normally I would take the Y-DNA Warehouse as the most reasonably-neutral comparison numbers though. But it would be better if we had some statistics about numbers of SNPs called of different types from the various kinds of testing.

Anyway just based on that I wouldn't go anywhere lower than a 30x myself, but as I said there's a lot of ways of looking at this based on numbers from any one company.

I’ve always heard that 30x is the way to go but what do I know? At any rate, I’m not in the market anymore because I’m all tested out. So they’ll have to come up with something really great before I consider buying anything else. Perhaps, a long read test when the prices come down? What do you think?

[Dave-V]

I’ve always heard that 30x is the way to go but what do I know? At any rate, I’m not in the market anymore because I’m all tested out. So they’ll have to come up with something really great before I consider buying anything else. Perhaps, a long read test when the prices come down? What do you think?

Personally I think either Oxford's nanopore or Pacific Biosciences' SMRT testing approach is the next wave of "long read" testing that will break through the current 23.6M base pair limit and read the rest of the Y chromosome. Nanopore has the edge right now since it was already used to sequence the centromere at least once and Dante at least already has a nanopore-based test on the market. But both are still too error-prone to be reliable enough for Y-DNA testing as far as I know. It may not be a cost issue; there is already a hand-held nanopore device available so cost may not be a factor. Error rates are certainly still too high though.

Full Genome's "Long Read" Chromium offering is really a short-read NGS with the addition of a pretty cool technique called linked-read sequencing that vastly reduces the number of misalignments and contributes to much higher accuracy in calling SNPs; that might be an interim option if they can fix the current $2900 USD price tag.

But no long-read improvement is really going to help much if we can't get a reference genome that covers more of the Y chromosome; even hg38 doesn't even cover the entire 23.6M base pairs that currently can be read. That's a current limitation even of the WGS testing that pushes "short-read" to its limits; we can read sections of the Y that we don't even know what the ancestral state of a base pair position is supposed to be. So until widespread testing helps us determine that, or the reference genome is expanded to cover more of the Y, we're limited on many positions of the Y that even if we can read them we can't "call" SNPs there. Most of the potential of "long-read" is in that category so there is still a lot of work to do.

[MacUalraig]

YFull tend to overstate the depth you need IMHO, the YSEQ WGSx15 gives me enough Y depth (~7x). YF will take the data if you line their palm with silver of course. I've submitted about 30 x15 kits and offhand don't believe I can recall any SNP calling issues which bothered me. The Data warehouse page is just more stats, yes they might be interesting to persue but I'm not convinced how meaningful they are in terms of our day to day hobby. Measures of false positives and negatives might help more.

Of course if you think depth is a bigger issue than I do they there are other market options to explore at 30x (Dante, Nebula, Sano) some of whom will work out cheaper at least if you time the order right. I've decided to largely stick with YSEQ for their post sequencing analysis and reports which beats everyone else plus their very good turnaround.