Why more sequences per sample with 454 in comparison with clone libraries?

[Sam299]

Hi,

Sources tell me with pyrosequencing you can get way more sequences per sample than with traditional clone libraries.

Is this because the Emulsion PCR is exclusively and inherent to pyrosequencing, and creates more amplicons than more traditional PCR amplification techniques?

Or is the reason simply because with pyrosequencing you can sequence up to a million reads in one run?

If the latter is true, am I correct in saying that you can get as much sequences per sample with traditional clone libraries, but analyzing one millions sequences in this manner would be practically impossible due to time issues.

Or are both wrong and is something else in play here such as something in the cloning and clone picking process?


Any help with this interesting topic would be greatly appreciated!

[SNPsaurus]

Picking a clone library will limit to the complexity to what you can manage to pick. It is a lot of work, even with automation, to pick a million clones. With next-gen sequencing, the "picking" step is replaced by emulsion PCR or clonal amplification on the flow cell (Illumina), so the complexity is instead bottlenecked by the # of sequences (millions with 454, hundreds of millions with other platforms).

The informatics for next-gen sequencers are designed to deal with millions of sequencers, so most applications will not require extensive time for analysis.

[Sam299]

Thank you for your answer!

But I still don't understand why pyrosequencing results in more sequences per sample...

Or is what your saying that emulsion PCR is inherent to pyrosequencing? And emulsion PCR is the reason pyrosequencing results in more sequences per sample because also the less abundant organisms get sequenced?

Then I wonder why it is not possible to see the less abundant species with clone libraries?

[SNPsaurus]

I'm not completely sure I am understanding exactly your question, but I'll answer anyway. Let's imagine you have a sample (a complex metagenomic sample). You divide it in half and send one half to a facility to create a clone library and sequence each clone by Sanger sequencing. The other half you send to a 454 facility.

The clone facility will fragment the DNA, clone into a vector, transform into bacteria and then pick 100,000 clones. These clones are Sanger sequenced.

The 454 facility will carry out 454 library prep on the extracted genomic DNA, and then sequence one million of the fragments in a run.

In the end, you get very similar data. The clone library gets you 100,000 sequences that corresponds to 100,000 genomic fragments. The 454 run gets you 1,000,000 sequences that correspond to 1,000,000 genomic fragments. So of course if it is a complex sample, there are rare species you will not see in the clone library that you will see in the 454 run, because you have sampled many more of the fragments in the 454 run.

[krobison]

It's also worth noting that the 454 and clone picking will show different biases. There really are regions which drop out small insert clone libraries. Any PCR-based sequencing technology has difficulties with extremely G+C rich or poor sequences.