LV Lib L seq run too many dot and mixed?

[chitownseq]

Hi everyone,

Tried my first LV run last week and I ended up with a higher percentage of dot and mixed than quality reads.

Does anyone know what can cause this?

Thanks!

[RCJK]

Hi there, this might help:

From Part B of the software manual:

• Dots Filter – fails for reads with too many negative flows. This filter discards reads that are too short or have too many poor incorporations or interruptions. A ‘dot’ is an instance of three successive negative flows (no signal for three base flows, denoted at ‘N’). Discards reads with the last positive flow occurring at less than 84 flows (~30-50 bp). Also discards reads with more than 5% of the flows before the last positive flow occurring as dots. (numDotFailed metric)
• Mixed Filter – fails for reads with too many positive flows. This filter discards reads with too many nucleotide incorporations, possibly occurring from a bead carrying two or more DNA fragments attached, a well containing more than one DNA bead, signal contamination from a neighboring well or a low signal-to-noise ratio well. Discards a read if more than 70% of the flows occurring before the last positive flow recorded a positive signal. Additionally, if the normalized flow signal value is between 0.45 and 0.75, it is considered ambiguous. If the read has more than 20% ambiguous or less than 30% unambiguous positive flows (either below 0.45 or above 0.75) the read is discarded. (numMixedFailed metric)

[chitownseq]

Thank you for your reply.

Yes I read that in the analysis manual too. My question is how do you prevent getting more than one DNA fragment attached to a bead? Or how does more than one bead go into a well? (didn't think that could happen so easily) and how do you improve a low signal to noise ratio?

I'm trying to figure out exactly what I did that could cause any of these things since my amplicon runs have been successful.

I suppose I am assuming that a high dot and mixed % is operator error. I just can't figure out where this operator erred!

[RCJK]

What was your % enriched beads recovered from the LV emPCR? A high recovery can indicate a likelihood of there being >1 template per bead. Adjusting the cpb of input DNA so that the % enriched beads are in the range they recommend should help with this.

I'm not really too sure about the high dots though. I'd guess some of it could be if there were any adaptors or other small fragments that were carried over into the emPCR, or also maybe something related to the template and possibly its quality??

(Sorry about the Bears, they should've won!)

[chitownseq]

In region 1 I had 5.3% enrichment and region 2 I had 5.5% enrichment.
I calculate my % enrichment from total input beads.

My enrichments were so low I almost didn't sequence. The last thing I was worried about was putting in too many copies per bead! This is why I am so confused.

I only put in .15 copies per bead as per nimblegen's instructions from their sequence capture protocol. I also ran an agilent of my library after sequence capture to check size and ran a pico green to quantitate.

Thank you again for your help.

(We had a huge party today to watch the game as we are crazy Bear fans and we are all very sad that they lost..... Especially to those Packers grrrrr.)

[Soulbee]

did you perform a titration with SV oils? this could give you a better idea for cpb and you dont have to do a four point titration, just do two - one really low cpb and a higher one.
How long are your fragments you are sequencing? if they are short, they can be over-amplified during emPCR and give off a very strong signal during sequencing which would give you bleed over signals into the neighboring wells. This may cause high dot/mix. Maybe try loading a bit less beads to give better resolution? and make sure you check for broken wells before breaking and enriching (settled white bead at the bottom of the wells = broken; hold your plate up to a good light and look at your plate from the bottom, the light should help in detecting settled pellet of beads at the bottom of the wells) and take them out.

[anujgupta]

Agree with Soulbee. I wish to add one more thing that while loading PTP at the time of sequencing run, Do not try to suck and push by pipetting in order to remove bubbles. Pipetting should be in one go and smooth. To avoid bubbles you can tilt PTP little towards you, raising from other end. Bubbles and suck and push by pipetting cause higher dots in the result.

[chitownseq]

Thanks.

I did run a titration with SV. .15 copies per bead gave me 9.6% enrichment. I used .15 cpb in my LV.

I don't think we have ever seen a broken emulsion in our lab (and we do look every time) but maybe we can look closer at our plates when they are done cycling.

My size range is 300bp - 1000bp. My average length sequenced was 350bp.

Roche had advised us to do another double SPRI clean up on our library in order to make sure all small fragments are gone. We did do the double SPRI, ran another agilent etc. When setting up LV to sequence we over enriched (and now Roche is trying to figure out why)

Thanks for your help/suggestions!

[MissDNA]

If your library is fine, your enrichment was low, you might have an instrument problem.
Today we had a failed run. 99% of the reads were filtered on dot+mixed. Right away I sent all the data to Roche. Turns out we may have some instrument problem. They asked us to perform 2 VolVers. The results were ok but they asked us to stop using the machine till an engineer checks it.
Every time I have a problem with a run and I know my sample was fine, I report to Roche.

[Soulbee]

After you did your second size selection, did you confirm you size? I have had times where SV and LV oils do not correlate with each other as far as cpb goes.. which made the titration pointless. After sequencing, how were your control beads? If you got nice sequencing results for the control beads, it is not the instrument. High enrichment and/or high dot/mix is most likely due to either broken emulsions or not washing out the white beads properly during enrichment. Sometimes, some samples I have enriched required very many washes (~20-30) and found that you can't just stop at 10 as the protocol suggests. Once you think you've done washing, get a clean tube, transfer your last wash supernatant into it and spin it down to see how much white beads are still coming off. there will always be a little bit but should be almost invisible.