[ETHANol]

I don't know if I would say your Bioanalyzer image is inaccurate. Your agarose gel is overloaded and not run very far which will result in larger fragments running faster then they should. Your average fragment size appears to be 600 bp by intensity which means on a molar scale it would be less. The smaller inserts probably cluster more efficiently (I am just guessing here). So no in your case I do not think the agarose gel is more accurate then the Bioanalyzer.

Another explanation is you may have some over amplification issues as well, which result in the fragments running bigger then they are. If you combine this with the overloaded agarose gel making them run slower then they are, it makes the agarose gel look more accurate. The end result is the two artifacts canceling each other out. While ok for you in this case, it's not something I would want to rely on.

[shanshanzhou]

Hi All,

Some updates from me. I followed ETHANol's protocol after tissue grinding and sheared the chromatin using the same settings (5%d, 4i, 200cpb) for 6 min. I seem to get some improve at least for one of the antibody I used (H3K4me3). The peak still centered around 1000bp but is a lot broader than before. See attachment (A-4 is prechip DNA sheared 4 min, A-6 is prechip DNA sheared 6min, K27 A-6 is postchip DNA for H3K27me3 and K4 A-6 is postchip DNA for H3K4me3)

Any suggestion on improving more? I'm thinking to decrease the shearing intensity and increase the shearing time. Has anyone done that before? If I want to do so, shall I drop the intensity parameter only, or I need to adjust Duty Cycle and Cycles per Bust as well?

Best,
Shanshan

[tamara.tjitrowirjo]

Dear ETHANol,

I do not think that my agarose was overloaded, I was wondering how could you judge that since the figure that I uploaded was already gel-extracted.

And assuming that your hypothesis is correct that smaller fragments cluster better, should I not see a wide range of fragment length instead of a a very tight distribution with a median of 150bp (sd 50)? We have previously sequenced fragments all the way to 800bp. This is only to say that longer fragments (at least until 800bp in our experience) managed to cluster well.

Lastly, we found enrichment sites that others have found as well plus some that we validated in our qPCR as well. So I think, yes, this result is good enough for me. I have also read others posting about size discrepancy between bioanalyzer and agarose gel in the context of ChIP-seq sample prep, and when they sequenced they got similar result as I did, i.e. short fragments size contradicting the bioanalyzer result. I am aware that this problem is specific to ChIP DNA since I have also sequenced fragmented genomic DNA and found no such problem. Currently I am unable to explain this problem and I do not want to spend enormous time trying to fix this if I already got my result that is fine.

Thank you.

[ETHANol]

I totally agree that you should not spend more time trying to fix this problem. It's one of those things that is a little weird, but at the end of the experiment it doesn't matter.

From my experience, smears take some time to resolve to the correct size on agarose gels and the more DNA that is loaded in the gel the longer you have to run the gel to get your smear to resolved to the correct size. Because of volume restrictions it is hard to run a gel long enough if you do the amplification before size selection. With all that in mind I would trust the Bioanalyzer image more. Or it could also be the over amplification problem that I mentioned as well (in this case neither methods are accurate). But of course all this doesn't matter for your experiment but it is still nice know what is going on as it can be useful knowledge for optimization and trouble shooting.

[ALive]

Hi Everyone,

I've managed to get rid of the enormous fragments in my ChIPs. I'm not sure of what in particular did the trick... The main differences were:
- cross-linked for only 5 minutes
- Used a fancy new Bioruptor
- Determined the chromatin yield so I knew exactly what to put into my IPs (I used 5ug chromatin)

I've attached a photo of my agarose gel with the ChIPed DNA from two ChIPs (loaded 20 ng)

Cheers,

Ali

[Yzatis]

Hi everybody,
still having some issues regarding the discrepancy between agarose and bioanalyzer results on ChIP samples (fortunately no difference between input and IP). I've read with attention all your advices. My protocol is 10min Xlinking and 21 min bioruptor sonication. I tried decreasing X-linking to 8 min, as well as increasing sonication times, but the point is that in each case I loose my ChIP efficiency, and barely decrease the proportion of these 2kb fragments seen on Agilent and not on gel. So my best conditions for ChIP are the ones giving around 40% of fragments in the range of 2kb, while the rest is at around 250 bp.
My question is: is it really a bad idea to try to start the library with this, or should I try to shear my purified DNA after the ChIP (I can't extract on gel, I struggled a lot to get my 20 ng) ? I really don't know what to do...
Many thanks for your help,
Best regards,
Yzatis

[Jmab]

Thank you to Hamid. Same problem probably because over-crosslinking

Hi,

The problem with over cross linking of samples is that they will have to processed with higher amount of energy, and for a longer period of time to get the chromatin sheared to the smaller fragments. As a result most if not all of the epitope is destroyed during the process. You do get a very nice band showing great shearing, but the epitope will be destroyed. A great way to test it is to carry out a shearing time course, and then use 10-15ul of the lysates for each time point to do a western with the same antibody you will use for your IP. Use unsheared material as your control. You can easily then see the extent of damage to the epitope.
I have attached some data on the effects of over cross linking.
I hope it is helpful.

Thank you

hamid

I find just today this post. Thank you Hamid.
I am doig CHIP experiments from mouse liver. Following the described protocol for that tissue, I have problems to obtain correct size DNA fragments. When running on 1.5% agarose gel, I have always the same result, whatever be the sonication time/cycle/power conditions in a diagenode bath bioruptor. That is one band at 1700-2500 bp and another at 100-150 bp. No smear at 200-600 bp which is my aim!. One idea come from your forum posts: I am fixing the samples 10 min with Formaldehide 1% (SIGMA F8775); I've just see in the label that it contains 10-15% methanol
Do you thing it would be presumably the cause of incomplete sonication? Thanks!

[Hamid]

Hi,

There are several possible causes for what you are noticing:
1. over-fixation. We have processed quite a bit of liver tissue for ChIP, and we found that 5 minute fixation is quite sufficient for both histone and transcription factor ChIP. Also please use methanol-free formaldehyde to render your fixation more reproducible.
2.How are you preparing your tissue for chromatin shearing is also important. We have found that you should fix the tissue before you freeze fracture for nuclei preparation. Freeze fracturing prior to fixation releases quite a bit of chromatin which is then fixed more readily than intact chromatin in the nuclei.
3. Liver tissue is rich in RNA. RNase treat the sheared chromatin prior to proteinase K assisted reversing of the cross links at 65C.
4. if you are noticing 100-150bp fragments right away, then you are exposing the chromatin to too much energy. You should always look for a gradual reduction of size and distribution over time.

Thank you

Hamid

[Jmab]

Thank you Hamid

Hi,

I've been done some changes in my protocol, and, as Hamid suggest, I can say that:
1) the 1700-2000 bp band has disapeared by two means: a) fixing with methanol free- formaldehide and b) Quenching the crosslinking with fresh made glycine instead of using pre-made glycine stock -as suggested by a collegue from another lab-
2) I have reversed the crosslinking (heating the samples at 65º for 18h) an I've treated the samples with Rnase and ProteinaseK. Now appears clearly that the 150-200 bp band is DNA, probably too fragmented by excessive sonication.
3) Since, prior to consult Hamid post, I considered sonication was insuficient (because of the continous presence of 1700-2000 bp band in the gel), I sonicated the samples 35-37-40 cycles of 30sec/30sec, witch is really a too high energy.
4) I've repeated today the sonication, but in this case, with less cycles, and fixing the tissue with free-methanol formaldehide and quenching with fresh-made glycine. Tomorrow, I hope that the experiment conditions will be ready to do the CHIP.
Thank you very much,

jmab

[Jmab]

Answering the question of Hamid about the conditions of the tissue used in the experiments: I use fresh liver, wich is fixed and then I do the nuclei isolation protocol.
Thank you!
Jm

[pr393]

Hi Jmab where did you get the methanol free formaldehyde?

Hi,

I've been done some changes in my protocol, and, as Hamid suggest, I can say that:
1) the 1700-2000 bp band has disapeared by two means: a) fixing with methanol free- formaldehide and b) Quenching the crosslinking with fresh made glycine instead of using pre-made glycine stock -as suggested by a collegue from another lab-
2) I have reversed the crosslinking (heating the samples at 65º for 18h) an I've treated the samples with Rnase and ProteinaseK. Now appears clearly that the 150-200 bp band is DNA, probably too fragmented by excessive sonication.
3) Since, prior to consult Hamid post, I considered sonication was insuficient (because of the continous presence of 1700-2000 bp band in the gel), I sonicated the samples 35-37-40 cycles of 30sec/30sec, witch is really a too high energy.
4) I've repeated today the sonication, but in this case, with less cycles, and fixing the tissue with free-methanol formaldehide and quenching with fresh-made glycine. Tomorrow, I hope that the experiment conditions will be ready to do the CHIP.
Thank you very much,

jmab

[Jmab]

Hi pr393,
Here the product and catalog number:
16% Formaldehide (Methanol free)
Thermo #28908

http://www.piercenet.com/product/formaldehyde-ampules-methanol-free

Good Luck!
Jmab

[pr393]

Size shift in the Bioanalyzer problem solved

Hi pr393,
Here the product and catalog number:
16% Formaldehide (Methanol free)
Thermo #28908

http://www.piercenet.com/product/formaldehyde-ampules-methanol-free

Good Luck!
Jmab

Hello,

Thanks for the reference Jmab. Regarding this subject previously I was cross-linking my samples for 10min and then when it came to sonicating I really had to go for a large number of pulses (about 18 with 30sec on/off in the Bioanalyzer) without seeing significant changes in the size of the sonicated chromatin (always around 500-400bp and then about 1000bp in the Bioanalyzer/TapeStation, and after IP my fragments would be as big as 2000bp.

So reading the stuff posted here I've decided to optimize my cross-linking time and my the amount of sanitation pulses. Furthermore I've also used methanol free formaldehyde.

So far so good. Independent of the cross-linking time with very few pulses (minimum 10) I already get very decent bands in the gel and when I load the samples in the Bioanalyzer/TapeStation I no longer see a shift in the size of the band and I get some nice peaks which are a lot better than my previous experiments. I'm conducting an IP with this chromatin to check if different cross-linking times might have an influence in the final size of what I IP and if then I get fragments which are good enough size to do the Library. But it looks a lot better than before (see attached).

My conclusion so far is that the methanol in the formaldehyde is really complicating things, as pointed out by other users in this thread. But I can't still rule out over cross-linking.

Will keep you posted and thanks for the handy tips.

[Jmab]

Good work!

Hi prp393,
Well done! Your data seems promising. It is clear than Methanol free fromaldehide helps to decrease highest (2000bp or more) and to increase correct size bands.
Probably we need to find a balance between quality and time of crosslinking and sufficient ammount and quality of RNA in the correct size range.
Tell me if the result of ChIP reflects these good sonication results.
Thank you for your interesting reply!
Best!
Jmab

[hamzairshad267]

I am using ABI3500 for capillary fragment analysis, Can I just refill the cathode and anode buffer instead of changing them every time to make it cost effective?
Thankyou