fragmenting amplicons?

[greigite]

Hi,
I have a bunch of 1-1.5 kb amplicons that I'd like to run on Illumina. Has anyone had success nebulizing such small fragments? Or perhaps using DNAse or restriction digestion to bring them down in size a bit? Primer design is tricky for my particular targets so re-designing to get optimal size fragments is not always possible.

[ScottC]

Do you have access to a covaris instrument? I have sheared small fragments with an S2 before, and it has worked well enough.

Cheers,

Scott.

[greigite]

Unfortunately, no, I don't have a covaris instrument. Though if anyone knows of one of these in the general vicinity of San Francisco that is available for use, please speak up.

Do you have access to a covaris instrument? I have sheared small fragments with an S2 before, and it has worked well enough.

Cheers,

Scott.

[sci_guy]

While not as good as a Covaris S2 for this application do you have access to a Bioruptor? Just by wary that you'll see high coverage bias at the original amplicon ends. They will be available for ligation of the sequencing adaptors more often than other random sheared and repaired ends. You could try blocked primers to counter this. There is a paper on a scheme such as this published sometime recently... a bit vague I know , but its late here.

[greigite]

No Bioruptor either... we are old school out here

While not as good as a Covaris S2 for this application do you have access to a Bioruptor? Just by wary that you'll see high coverage bias at the original amplicon ends. They will be available for ligation of the sequencing adaptors more often than other random sheared and repaired ends. You could try blocked primers to counter this. There is a paper on a scheme such as this published sometime recently... a bit vague I know , but its late here.

[sci_guy]

One option to consider is concatamerisation of the PCR products before shearing with a nebuliser. In this respect you'll reduce the amplicon end representation bias considerably. On the other hand, it creates a pretty awful bioinformatics headache

We've tried this with Nimblegen SeqCap fragments coupled with the Illumina GAII. I wouldn't recommend it for SNP/INDEL discovery in cancer (my groups interest). For the rest of our samples we're skipping the concatamerisation and waiting for 100 bp GAII paired end reads. This will simplify adaptor removal informatically while still giving us plenty of sequence left in each read. If you're doing mutation discovery I'd recommend to only attempt concatamersation if you, or someone you have access to, is proficient at Perl/Python or the UNIX tools AWK/Grep.

[greigite]

Concatamerization is an interesting idea, but I think not suitable for our project. We are resequencing amplicons from a specific organism from a large natural population, and mapping reads back onto an assembled genome. My main concern with using concatemers for this is the potential incorrect mapping of chimeric reads. These could easily look like recombinants and obscure the presence of real recombinants in the population.
I'm still curious to know how chimeras could be untangled. It seems to me that the presence of concatemer chimeras could reduce the amount of reads that can be mapped back on to the assembly. Would you have to cycle through non-mapped reads, try to split them, then re-map? I am proficient in Perl so bioinformatics is not an issue.

One option to consider is concatamerisation of the PCR products before shearing with a nebuliser. In this respect you'll reduce the amplicon end representation bias considerably. On the other hand, it creates a pretty awful bioinformatics headache

We've tried this with Nimblegen SeqCap fragments coupled with the Illumina GAII. I wouldn't recommend it for SNP/INDEL discovery in cancer (my groups interest). For the rest of our samples we're skipping the concatamerisation and waiting for 100 bp GAII paired end reads. This will simplify adaptor removal informatically while still giving us plenty of sequence left in each read. If you're doing mutation discovery I'd recommend to only attempt concatamersation if you, or someone you have access to, is proficient at Perl/Python or the UNIX tools AWK/Grep.

[sci_guy]

My main concern with using concatemers for this is the potential incorrect mapping of chimeric reads. These could easily look like recombinants and obscure the presence of real recombinants in the population.

It's easier for Nimblegen SeqCap. In this instance there are known adaptors so you just need to search for them and remove. An adaptor can be present at the start of the read or two adaptors are present end on end completely (or partially) within the read. The ligation means adaptors A and B are present as one forward variant linked to the reverse variant of either A:A, B:B or B:A, A:B. There might be A's in between (from Taq) for the ends might be missing a few bases (exonucleases) so an approximate search is required. Fortunately Perl has String::Approx, which is a C wrapper for a C Levenshtein edit distance algorithm - so fast for Perl. This will run in O(kn) time, where k is the number of edits and n the length of the text (input). It's a still a big job. Really needs multi-threading.

Also, beware of sequencing through a PCR primer. There will a huge degree of masking of SNPs. Perhaps partial DNaseI or CviQI digests are worth considering?

[sci_guy]

You could also get fancy using adaptor containing PCR primers, melting the products and self ligating with circLigase. Then you can use rolling circle amp with Phi29. Create some really long DNA and shear this up. It's a bit technical. The Church lab has some IP in this area http://www.faqs.org/patents/app/20090099041

[Hamid]

Hi Greigite,

Please get in touch with me. The next time I am in the Bay area in for conducting demos of the Covaris Technology, I can stop by and provide you with a demo using your samples.

Thank you

Hamid

[JBuenrostro]

You should just use DNaseI. DNaseI is the most randomly fragmenting enzyme and has worked well enough in our lab. I would recommend finding the optimal DNase concentration by using a serial dilution of enzyme on a non-precious sample. Then repeating with your sample.

[pmiguel]

Hi,
I have a bunch of 1-1.5 kb amplicons that I'd like to run on Illumina. Has anyone had success nebulizing such small fragments? Or perhaps using DNAse or restriction digestion to bring them down in size a bit? Primer design is tricky for my particular targets so re-designing to get optimal size fragments is not always possible.

Presuming you do find a way to sonicate your amplicons, you may run into a end-bias issue. That is, we have seen at least one case where the vast majority of frags sequenced were from the very end of the amplicon--basically PCR primer sequence. This presumably results from reduced ligatability (even after end-repair) of the sonicator-torn ends vs. the PCR-fragment ends. So it may not be an issue if you use a DNase fragmentation method.

--
Phillip

[Lestag]

Frag and tag together

Hi, the Nextera system from EPICENTRE should be able to fragment and tag amplicons of that size or even smaller. Yes I have connections with the above co. but I believe this may a unique and good solution to the problem.

[pmiguel]

Hi, the Nextera system from EPICENTRE should be able to fragment and tag amplicons of that size or even smaller. Yes I have connections with the above co. but I believe this may a unique and good solution to the problem.

Hi Lestag,

Actually I posted a little about this in another thread:

http://seqanswers.com/forums/showthread.php?t=2878

One issue here: only amplicons that attract 2 Nextera transposon insertions would be suitable templates for PCR (bridge) amplification. Actually would require the insertion to be in opposite orientations to each other. Unless you guys have Y-adaptor ends in your transposomes.

--
Phillip

[Lestag]

two tags per frag

Hi Phiilip, yes, it's very probable to have two nonidentical tags, one at each end of the Nextera-generated fragment. Supression PCR makes it even easier to reach that goal. I've sequenced a handful of small, cloned "tagmented" genomic DNAs and the great majority of the PCR products have two distinct tagged ends. Amplicons of 1-1.5 kb should not be a challenge for Nextera, as we have gone as low as a few hundred bp with good results.