Poor R3 (35bp) mapping

[boscoliveira]

Is there anyone having problems with the F5 tags on Solid 4? Our two last lots gave us very poor mapping (28-50%), as compared to 75-85% for the F3 tags. ABI reps said that it could be a problem with the cleave solution, but a new lot, supposedly good, did not solve the problem. I wonder if this a localized or widespread problem.
Thanks.
Bosco

[dsidote]

We were told the same thing when we collected poor F5 paired-end data. We were sent new strip tubes for the F5 tag and a new lot of cleave 2.1 but the problem persists.

[boscoliveira]

Thanks for the reply. They checked our cleave 2.1 and said the lots were ok, the same for the F5 tags, but the data was just as bad. How bad is your F5 mapping?

[dsidote]

The F5 is only mapping about 40%. According to AB, 40-50% is the best they have observed for the F5 read.

[kopi-o]

We have had pretty low mapping rates for F5 tags, around 25-30%.

[westerman]

I'll agree that the F5 reads tend to be poor but since they are "free" we often run with them. I have seen lows near 20% and as high as 52% for F5 mapping.

R3 is a different matter. It should be within several percentage points of your F3 mapping. Our last F3/R3 run had mappings of 68.73% and 71.09% -- this was to a 10M year or so diverged species.

[boscoliveira]

Sorry, I meant F5 tags, not R3. So I guess our data is comparable to what you have seen.
Thanks.

[pmiguel]

I am pretty sure what we are talking about here are differences in the maturity of the 3' and the 5' ligation chemistry. I think the 5' chemistry is about where the 3' chemistry was with v2.

If you look at other threads on this topic, increasing numbers of FAM calls in later cycles are a complaint. But this would just be signal droop, right? As fewer templates strands are available for extension in each cycle, signal/bead drops to the point where residual background signal painting a bead gets called as the color. Speculation on my part, of course.

Anyway, mapping rates back in the 25-35 nt, v2 chemistry, tag length days was in the 20-50% range, wasn't it? I guess I could check.

Okay, I looked at an old chicken mate pair run with 25 bp tags from February 2009. That was v2 chemistry. We got 67% mapping with F3 reads and 71% with R3 reads. Chicken is a lot easier to map to than the human genome -- much lower repeat content because it is only 1 gigabase, compared to 3 gigabases for the human genome. Still, I would have to say that current 5' ligation (F5) chemistry is generating reads of a lower quality than seen for 3' ligation chemistry under v2. So it is not v2 quality, it is v1 quality.

--
Phillip

[SeqAA]

Pretty much ditto with pmiguel. I was told F5 needs some time to get up to speed with F3 in terms of mapping.