[captainentropy]

I think I need to create a new thread called "The Pippin Prep is dead! Long live the Pippin Prep!"

It is dead. In the lab next to mine (who actually owned it) it's fertilizing a back room full of old broken equipment. That's because I taught them this trick for size-selection that I learned and improved on. Note: this size-selection is post-PCR, which I've found to be vastly better for library construction.

1) Selection for fragments <600 bp
a. Raise the PCR volume to 100 μl with TE and add 65 μl SPRI beads (Beads/DNA = 65/100).
b. Vortex and incubate for 3 mins.
c. Collect beads on magnet and transfer the supernatant to a new tube.

2) Selection for fragments >200 bp.
a. Add 35 μl SPRI beads (This will change the ratio beads/DNA = 100/100)
b. Vortex and incubate for 3 mins.
c. Collect on magnet and discard the supernatant.
d. Wash the beads with 500 μl of 70% ethanol.
e. Dry beads for 5 mins @ RT.
f. Elute DNA in 25 μl EB.
1. save 1 μl for Qubit quantitation
g. Analyze on Bioanalyzer 2100.

Also, if you are certain that the fragments that you made your library from were in a range of 200-600bp then the first step (size-selection <600 bp) is not necessary.

Attached are Bioanalyzer traces showing the output of this method. The red trace is the discarded product from Step 2). The green trace is the recovered (retained on the beads) product from Step 1). The blue product is the fully size-selected library that was later sequenced. Notice that there is a lot of product (most, actually) that is < 600 bp. That's why I say that if the DNA used to make the library is in the range of 200-600 then there's no need for Step 1).

[Olaf Blue]

So, how close is this to the Broad/Illumina "Double-sided SPRI" method?

[TonyBrooks]

+1 for captainentropy's method. We use something similar (0.6X-0.8X Ampure XP) adapted from the NEBNext gel free method. Then PCR up and a final 1X Ampure XP cleanup.
Ampure <1X is actually pretty good at removing stuff below 200bp.
The standard Ampure protocol (1.8X) is good below 100bp, but so-so for fragments between 100 and 200bp - hence it tends to leave a lot of the 125bp dimer behind.

[Olaf Blue]

Thanks, Tony!

[captainentropy]

So, how close is this to the Broad/Illumina "Double-sided SPRI" method?

Yeah, after looking up their "Double-sided SPRI" method it looks to be basically the same. I was taught this technique by someone at the JGI. The "improvements" I made were: to dry the beads at RT rather than on a 50 C heat block (in the protocol I had), this made the beads easier to resuspend, and eliminate step 1 if the fragments were already in the correct size-range. After reading the Broad/Illumina protocol it looks like step 1 was made optional as well, for the same reason.

Also, and this wasn't written in the protocol, but I've used magnetic separators from Invitrogen and Diagenode, and with the Invitrogen rack I had to transfer the sample to a 1.5 mL tube but the Diagenode one allows me to keep everything in 8-well PCR strip tubes but the magnets aren't very strong, so I replaced them with some really strong magnets I got from K&J magnets. I can't say that it improved my library yield but the separations are faster now, plus I don't have any more of that iron-residue carryover I would usually get (would that stuff even interfere with any downstream steps?) with the weaker magnets.

[captainentropy]

FWIW, Beckman told me once they don't recommend using the Ampure XP beads for what we're using them for because the lot-to-lot variability is too great. I don't know about that, they seem to work for me. But, at the time, they told me they had a product specifically for the double size-selection in the works. It's available now as the "SPRIselect reagent kit."

[Olaf Blue]

Yeah, they told us the same thing when we called about this maybe 6 months ago (about "single-SPRI" or "Double SPRI" size selection). Told us that they have no apps data on this, so they can't recommend the beads for this. Hummmmm........has anyone used the "SpriSelect" Kit? Any reviews, both positive or negative?

[hawaii454-0]

amount of starting material

I recovered well over 2 ug of amplified library for each sample. Next time I will use this to my advantage and reduce the number of PCR cycles.

How much input DNA was there?

How many PCR cycles would you recommend for a standard TruSeq DNA library prep?

[captainentropy]

How much input DNA was there?

How many PCR cycles would you recommend for a standard TruSeq DNA library prep?

Input into the PCR or the library prep? Using the old-school gel purification method I would lose most of my DNA before the PCR. It was usually barely detectable I started using 25 ng, 50 ng, or more of DNA as input for the prep itself (not easy to always get that amount of ChIPed DNA). The Pippin Prep allowed for better recovery so I began using less DNA in the prep. But once I moved the PCR step prior to selection I wound up with a lot more DNA than with the preceding iterations of my protocol (gel purification, then PCR). So, now I only use 10 ng of DNA as starting material. This usually results in around 5-6 ng of DNA going into the PCR. I typically recover 150-200 ng of DNA after the PCR step now.

When I first started doing ChIP-seq I followed what Illumina recommended (and what other labs around me were using), so that would be 18 cycles. Now we use 10-12.

I now quantitate after the End-repair, Adapter-ligation, PCR, and Ampure purification steps. If there is a lot of loss during one of these steps and I have less than 3 ng going into the PCR step I'll use 12 cycles. Otherwise I use 10. I typically recover between 60 and 125 ng of DNA after size-selection. My protocol now is super consistent. Our sequencing lab knows exactly what to expect with my libraries now and can push our clustering density to 230 million or more.

[cbird]

Captainentropy,

We've been working on size selection in my lab.

SpriSelect works much better than the ampure xp beads for size selection. I'm not sure why this is. I'm trying to track down the answer. We make our own "ampure" beads for half the cost and I'd like to make our own "SpriSelect" also. It's probably just a difference in the PEG NaCl solution the beads are in.

On the gel extraction front. I too got poor results at first. Step 1, use blue light transilluminator. Step 2, the pH of the binding solution + melted gel needs to be below 7. We use pH strips, so it's not an exact science, but ~pH 6 yields almost 100% recovery of DNA. Don't trust the Qiagen colormetric test if you're using that. It doesn't indicate optimal binding pH.

I know people using the Flash gel system (Lonza) and the Pippin Prep. I'm skeptical whether these are worth it. The E Gel system seems to be the cheapest of those alternatives, with a running cost of $2.14 per sample.

Ultimately, the beads are the cheapest alternative, recover the most DNA, and can be automated.

[captainentropy]

cbird, I agree that the beads are the best choice right now. Beckman told me a while back that these (SPRI select) beads have lower lot-to-lot variability. Maybe there's an improvement in formulation too, I don't know. We recently switched to the SPRI as well because we finally used up our stock of Ampure XP (!). Almost 300 libraries we've made with Ampure and over 100 with the Pippin. Now we've made about 20 with the SPRI Select and I can't say I see a difference at all compared to the XP. I'm using it just as I did the XP (volumes, washing, etc.).

We're going to try the homemade version of the beads too. But the XP/SPRI-Select is about $6.80 per library for us, which isn't that bad. So, what is it in dollars per library for the homemade reagents you use? I'm a big proponent of DIY. I saved my lab in grad school about $10K alone by making ECL reagent (!). Science is expensive enough as it is :`(

My first few libraries were with standard gels and they sucked. We looked into the E-gel and it definitely would not work for us. We switched to the blue-light illuminator and SYBR-dyes. Longer gels, overnight runs in the cold room, we tried lots of things and the results were ok but once we started using the Pippin, that's when our ChIP-seq really improved. But we only use the beads now. (sorry Sage Science)

[pmiguel]

On the gel extraction front. I too got poor results at first. Step 1, use blue light transilluminator. Step 2, the pH of the binding solution + melted gel needs to be below 7. We use pH strips, so it's not an exact science, but ~pH 6 yields almost 100% recovery of DNA. Don't trust the Qiagen colormetric test if you're using that. It doesn't indicate optimal binding pH.

Interesting.

I would add a trick that we got from an FAS at Applied Biosystems -- don't heat the gel slices during slice "melting". Instead add more more of the chaotrope solution (QC?) to allow the slice to melt at room temp.

Much higher yields this way. (Although it may well be the pH issue you describe that governs yield. The chaotrope solution probably has a buffer in it -- the more you use, the better pH would be controlled...)

But also, even at 50 oC you might be melting strands of DNA that are AT-rich. Why risk it?

I am 100% with you on the blue light trans-illuminators.
--
Phillip

[saida]

Pippin Prep results

Hi all,
Could you please, let me to solve our problem. We have Pippin Prep and I want to size select my DNA library. I used two library and set 200 bp target (range 150-250) for the first and 350 bp target (range 100-600 bp) for the second one. At the end, I had enough concentration of DNA, but I have doubt if they are my targets or not, because I did not see any peaks on my sample lines. After appearing of 150 bp marker peak, elution started and the graph line dropped on both sample lines. My question is that, if that is normal? Do I have to see peaks corresponding to my targets or selected ranges on sample lines as similar to markers or no?
I am attaching screenshot.

Thank you in advance,

[nucacidhunter]

Screen shot looks normal. Lane 2 started elution at 150 bp and lane 3 at 100 bp. If you want to check the eluted libraries size distribution you need to run them on BA or similar device.

Edit: I wonder if the libraries were loaded without clean-up?

[saida]

Thank you for your reply,

We did clean-up before and after size selection using GenoCruz™ DNA Purification Kit. I was expecting ethidium bromide fronts due to elution (Brighter bands and small peaks) after starting of elution as shown in a next screenshot (attached). But I had different. Is that normal?

Unfortunately we do not have BioAnalyser. ((