Sequencing of bisulfite treated DNA - (Jul/25/2006 )
I have recently made an attempt to try bisulfite sequencing, and thanks to the huge amounts of advice here I am getting really nice PCR products (of the appropriate size approx 400bp). I have cloned these using the TA system and am now trying to sequence the inserts.
Most of the sequences that I have been getting back are REALLY poor quality! The ones that are good are good enough to determine that a) my PCR fragment is the fragment that I want and that the bisulfite conversion seems to be working well.
I have tried two different sequencing facilities (with them doing the reactions) with similar results. Are there any tricks to getting good sequencing results from these (rather weird) sequences? Some of the problems (although by no means all) appear to be the algorithms used for calling bases, but there also seem to be problems in the separation of the peaks.
Has anyone else had this problem? Are the sequncing facilities having problems or do I need to get them to adjust something?
The DNA I am sending them is clean A260/280 ratio close to 1.8 and I'm measuring the quantity using a nano drop.
Is there any difference in which universal primer I use as I have only tried T7 (the strandedness doesn't seem to make a difference with these clones).
Directly sequencing bisulfite PCR product is very challenging, but once your PCR product is in a vector, you should be able to get pretty nice sequencing results no matter there is methylation or not. You can just use standard sequencing primer specified by the vector.
Have you verified that every clone contains your insert? We usually do a colony PCR and then get overnight culture only from those positive colonies.
Check with your sequencing service provider to optimize the sequencing reaction by adjusting the amount of DNA and primer used.
I agree with pcrman,
once in a sequencing vector, it should be quite easy to get some nice sequencing results.
We verify our inserts with a simple enzyme digest after mini-prepping prior to sequencing.
I have switched to mainly using SP6 as the sequencing primer as it seems to work more consistently than T7 in my hands. Maybe give this a go, M13(+) or (-) seem to work well. But as pcrman has said, once in a sequencing vector you should get some nice results out from it.
That's what I would have thought (yes I did check for the correct insert). I just wanted to make sure. Might have to go and harass the sequencing folks again.....and I thought the actuall sequencing was going to be the easy part!
What sequencer are you using? There certainly could be issues which depend on the sequencer, since the sequence has essentially no C's (or, in reverse, essentially no G's). The gain control for the dye detection in the sequencer could try to compensate for this, and make very bad mistakes. I have not had this problem on the ABI 310 or 3730, but can't speak to other platforms, and could certainly believe it exists.
you are correct, there are issues about the basecalling software of the ABI's that over compensate for the base composition bias however it is exemplified when directly sequencing PCR products and is another reason why it is a challenge to directly sequence products.
once you have the vector with insert it's very easy to sequence and your base biases are less pronounced becuase you would most likely sequence parts of the vector before and after the insert. I haven't had any problem with sequencing through the insert nor this compensation whilst sequencing vectors.
I had similar problems several years ago when I was doing manual sequencing with 33P labelled primers. The sequencing of the vector part was perfect however the sequence of the insert was terrible. After ~100 bp from the begining of the insert Ts started to disappear and the C lane was "dirty" with nonspecific bands. I was able to read olly about 200 bp from the primer. With the automated sequencing the situation was similar.
The most striking thing about all of this is that the sequences of some clones were perfect!
Maybe these problems were related with the secondary structure of the bisulfite modified region.
I just kept sequencing until had enough reliable data. it was really reall much work to do!
Maybe it is worth trying different cloning vector, with different sequencing primers? And also the annealing temperature during sequencing reaction.
I also tryed to add more dTTP into the MixT and it helped a little. However it was with the radioactive sequencing.
I found t7 reading 5'->3' to be a sloppy primer with p-Gem vectors, I switched to sp6 3'->5' for sequencing and it has been golden.