Reason for max read length?

[trauba]

What exactly is it that limits the read length anyways?

Also, eventually some oligos get "out of phase" and then muddy the signal. Is this somewhat caused by insufficient purging of reagents from previous cycles?

Can anyone elaborate on this? Would be greatly appreciated!

Thanks!
Jim

[GenoMax]

For what kind of technology? For illumina probably reagent stability. One could get 620 bp SE reads by using a 2x300 PE kit and setting up an asymmetric run on MiSeq.

Read about the phasing issues here. Rather than the purging of reagents phasing issues are likely due to non-efficient unblocking and/or incorporation of multiple nucleotides during one cycle.

[Markiyan]

Depends which technology we are talking about.

There are both technology dependant and independent factors.

Technology independent:
1. Input DNA/RNA length.
2. Input DNA/RNA damage/modification (Nicks, Base Mods, cross-links, etc in addition to 1), depending on methods/enzymes used during sequencing.
+ Input Library size distribution (for methods requiring it)/Ligation efficiency drop for long fragments in some methods.

Technology dependant.
1. For Single molleculle based methods:
- Limit of the enzyme processivity/stability (Helicos/PacBio)
- Physical constrains during relocation of the long stranded DNA/RNA structures in the pore (Nanopore).

2. For PCR-based methods (Illumina, 454, iontorrent).
- Limit of the PCR efficiency over the long products/secondary structures/extreme AT/GC% and homopolymers.
- Dephasing during sequencing reaction (caused by loss of synchronisation due to processivity, missed incorporations, polymerase slippage etc events.

3. For linear PCR-amplicon based methods (Sanger/Maxman Gilbert) - Loss of the electrophoresis resolution of the larger DNA fragments due to gel matrix separation limitation caused by diffusion, DNA structure and other factors.

PS: Main Hinder for Illumina's longer reads now is dephasing, which is also very sensitive to reagent quality and has an exponentially increasing effects on error rates as the reads length increase.

PPS: even a completely random read can be considered to be 25% accurate if the expected GC=50% :-)