[MissDNA]

I think Roche decided to do the first upgrade here in Brazil in that other lab, because they have two GS-FLX, which is not our case.
It was indeed some software problem and now they are waiting for some installation CD, in order to complete the process. Since things take a lot longer around here, it will take awhile.
Biggest problem is we already ordered 7 FLX+ kits, and they are supposed to arrive in January...

For the people who already have the upgrade: are the runs using Titanium still good?

[eladSeq]

Amplicon sequencing not supported on FLX+

Thank you everybody for your input.

Indeed, FLX+ is not currently supported for amplicons. From my understanding, there would be too much "flashing" due to the homogeneity of the sequences, such that SNPs would get obscured.

So I'm planning to use the regular titanium chemistry and split my gene into 3 amplicons, that can be covered by the 400 base reads.

I was informed that adding multiple (10 or more) different barcodes for each amplicon would help add complexity to the mixture such that bases are "offset" in the flowgram and reduce amount of "flashing". Has anybody tried this or recommend it? When I spoke to Roche rep he said that didn't have any data about it, and that it wouldn't make a difference if using MIDs since they are all the same length. However, it seems to me that the sequence matters, since the nucleotides are added in a specific order.

For example, supposing bases are added in the following order: T A C G
A barcode of CTG would take 8 nucleotide flows to complete, while a barcode of GAC would take 7 nucleotide flows, thereby offsetting the start of the desired sequence.

Thanks,
Elad

[ajthomas]

That's an intriguing idea. It seems to me that should help. If you're going to do it, you should carefully pick your MIDs so as to maximize the effect. After reading this, curiosity got the better of me and I looked at the MIDs I use to see how much effect they may have. I use the 48 MIDs that Fluidigm recommends for their Access Array system, which are mostly just Roche's first 50 MIDs with a few replaced that I assume they found didn't work well for one reason or another. Here's what I found:

flows # of MIDs
14 1
15 3
16 6
17 22
18 9
19 3
20 4

Nearly half of them require will end on the same flow, so if you try this route, be selective in which MIDs you choose.

[eladSeq]

So I have designed sets of A and B fusion primers with varying barcodes to get the desired offsetting of the flow reads. This will require 42 primers to make and pool 21 amplicons. Because of the number of primers needed, we have decided to go with ordering regular desalted oligos rather than spending a lot of money on purified. I have come across some researchers who use desalted and some that use purified, but no side-by-side comparison whether there is a difference.

Some brainstorming on the effect of oligo purity and questions I have about the emPCR steps in general:

At this oligo length (54-59 bases) IDT says that about 65% of the synthesized oligos will be full length. So would the 5' truncated products cause a problem? The PCRs should work fine since the 3' ends would be good. Perhaps 5' truncated amplicons would bind less efficiently to the capture beads, but the titration step should account for this and determine the optimal DNA loading quantity, right?

Would there be an issue with the sequencing primer annealing? I'm not sure how this part works. Does the sequencing primer anneal to the A or B adaptor sequence?

Also, I am unclear as to how the bead enrichment step works and what this actually does, if anybody could help me understand this?

Sorry about all the questions, but I am hoping to try and get this library prepared correctly the first time, and avoid problems down the road. Thanks all.

[ajthomas]

The truncated oligos might or might not bind to the beads or the amplification primer, depending on where they're truncated. Where this might be a problem is if the primer with one MID has significantly more or fewer truncated oligos than the others. In that case, the balance of reads you get out may be a little off from what you desire and put into the system. But yes, once the molecule binds and goes through the first round or two of amplification, the molecules will be full length. So as long as a particular molecule isn't truncated at one end or the other far enough from the end that it won't bind, it will be fine.

By the time you add the sequencing primer, the molecules will be full length, so that won't be a problem.

The bead enrichment step works by attaching a biotinylated oligo that matches the adapter that is not attached to the bead. The biotin is used to separate the DNA-containing beads from those with no DNA (hence "enrichment"). Ideally, you only want about 5-20% of the beads or so to have DNA on them, because if you have more than that you will have a lot of mixed reads.