New to Next Gen

[abhagwat]

Hi! We are just starting out with Next Gen sequencing and I see no easy path to developing a project. It seems that most genomics centers are stratified along platforms, some mostly do Next Gen sequencing while others microarrays. And some don't do the bioinformatic analysis. Is this a correct perception? Would be nice to have advice on how to proceed.

[ECO]

What are you trying to do? No one can answer your broad question as phrased.

[adaptivegenome]

Good starting point:

http://www.ploscompbiol.org/article/info:doi%2F10.1371%2Fjournal.pcbi.1002202

[abhagwat]

We are trying to map regions of the human genome that contain a certain DNA modification. We have developed a method to pull down DNA containing such adducts and would like to determine where these pieces come from. Also, we are trying to correlate the location information to transcription activities in those regions. The RNA part is relatively straightforward as there are a lot of "expression chips" available. The DNA part is confusing because whole genome NextGen sequencing may be an overkill here. Most likely the adducts are concentrated in only 1-3% of the genome. I am afraid of being deluged with a lot of useless information which will take huge amount of bioinformatic effort to weed through. We considered oligo arrays (ENCODE and tiling), but they have their own disadvantages. ENCODE arrays cover too little of the genome while tiling arrays cover too much again creating the same problem as NextGen sequencing. My other problem so far is that most University centers seem to do NexGen sequencing over ChIP-on-chip and other assays and do not always provide good software support. My general feeling is that they are just technicians without a detailed understanding of what they are doing. I have been calling up chip manufacturers, sequencing service providers as well as genomics center operators for the past two weeks and getting nowhere.

[adaptivegenome]

We are trying to map regions of the human genome that contain a certain DNA modification. We have developed a method to pull down DNA containing such adducts and would like to determine where these pieces come from. Also, we are trying to correlate the location information to transcription activities in those regions. The RNA part is relatively straightforward as there are a lot of "expression chips" available. The DNA part is confusing because whole genome NextGen sequencing may be an overkill here. Most likely the adducts are concentrated in only 1-3% of the genome. I am afraid of being deluged with a lot of useless information which will take huge amount of bioinformatic effort to weed through. We considered oligo arrays (ENCODE and tiling), but they have their own disadvantages. ENCODE arrays cover too little of the genome while tiling arrays cover too much again creating the same problem as NextGen sequencing. My other problem so far is that most University centers seem to do NexGen sequencing over ChIP-on-chip and other assays and do not always provide good software support. My general feeling is that they are just technicians without a detailed understanding of what they are doing. I have been calling up chip manufacturers, sequencing service providers as well as genomics center operators for the past two weeks and getting nowhere.

This makes lots more sense. It is a shame you have had bad experiences with genome centers. I trained at the BCM genome center and it was a great experience. We frequently developed new techniques, etc. The bioinformatics was top-notch...

I think that nextgen can help you as your project is a variation of capture sequencing. Here are some questions though before we can figure if it will help:

1. How long are the DNA fragments you pull down?
2. How many are there? How much actual nucleotide sequence will you get in the end.

Knowing these two things can help you pick a platform and the bioinformatics for what you want to do is straightforward. If there is not someone nearby you then you can PM me and I can probably help.

The alternative if you have not so much DNA and short fragments is to just clone them into a vector and sequence by Sanger.

[abhagwat]

This makes lots more sense. It is a shame you have had bad experiences with genome centers. I trained at the BCM genome center and it was a great experience. We frequently developed new techniques, etc. The bioinformatics was top-notch...

I think that nextgen can help you as your project is a variation of capture sequencing. Here are some questions though before we can figure if it will help:

1. How long are the DNA fragments you pull down?
2. How many are there? How much actual nucleotide sequence will you get in the end.

Knowing these two things can help you pick a platform and the bioinformatics for what you want to do is straightforward. If there is not someone nearby you then you can PM me and I can probably help.

The alternative if you have not so much DNA and short fragments is to just clone them into a vector and sequence by Sanger.

Ans.
1. The expected average size is ~250 bp (4 bp restriction enzyme), but we can make it a bit longer.
2. Don't know how many fragments there are. On an agarose gel it's a smear.

[adaptivegenome]

So I would first start with isolating the smear (I assume you would run it off the gel or do a gel extraction). There is a minimum amount of DNA necessary for sequencing on nextgen instruments. Then I think for 1%-3% of the genome I would vote for an Illumina GAII run (HiSeq is overkill) or maybe Ion Torrent. Depends on what you have available to you.

If you do not have the minimum DNA and even if you do, I would probably start with the easy experiment -- add some A/T overhangs to your product, TOPO clone, and Sanger sequence. See if you are getting sequences you expect and if it is useful. Perhaps this will be enough data. If not it makes a good validation set to confirm things are working when you move to a high throughput sequencing method.

I hope this helps. Feel free to ask more specific questions about any of the steps... (and the above is just my thoughts; I'm sure others might have more ideas for you)

[ETHANol]

Yes, the more time you spend construction your post and more specific your questions are the better the answer you will get here. And if your question is well written you will get some good answers.

As far as the wet-lab side, assuming you will use the HiSeq I would take a look at my ChIP-seq and MeDIP-seq protocols. If you have to denature the DNA prior to enrichment for you DNA marker then go with the MeDIP-seq protocol. If you do the DNA marker enrichment on double stranded DNA then follow the ChIP-seq library protocol. You would have to increase or decrease the adapter concentration to the amount of DNA you have but that's about it.

1-3% of the genome is a lot of the genome so NGS is not overkill especially in that you can barcode essentially as many samples as you like and fit them into one lane on the HiSeq.

The protocols:

http://ethanomics.wordpress.com/medip-seq-protocol-with-truseq-adapters/

http://ethanomics.wordpress.com/chip-seq-library-construction-using-the-illumina-truseq-adapters/

[NextGenSeq]

Note be careful using restriction enzymes in next gen sequencing. These create libraries which have the same sequence motif which means when they are sequenced all the clusters have the same sequence in that region of the library. This leads to difficulties in resolving the clusters and poor sequence quality in those regions.