HiSeq vs MiSeq read-length and quality

[weehzer]

Hi,
I have a very basic question.
Why is the read highest achievable read-length of MiSeq compared to HiSeq longer although, as far as I understand, they use the same chemistry?

Second question would be, why is it usual that the second read of a paired end read has lower quality than the first one?

I am a novice in the NGS field. I searched the forum but couldn't find an answer.
Are there any comprehensive introduction to illumina sequencing technology?
The manuals on the illumina webpage are nice, but only give a very rough overview.

Thanks a lot.

kind regards

[GenoMax]

MiSeq - 2 x 300 bp (or 600 bp single-end, theoretically possible though you probably don't want that).
HiSeq 2500 - Max supported read length is 2 x 250 bp in Rapid-mode.

R2 generally has lower quality as:
a) cluster grow fatter over time and the sequencer may have trouble basecalling if a run is overloaded
b) reagents are held at 4-6 C in sequencer but deteriorate some over length of the run

[weehzer]

Thanks a lot, very nice explanations!

[bbp]

My understanding that the read lengths are limited due to DNA damage.

Firstly, incorporation chemistry/removal of the dye terminator is somewhat corrosive to the DNA - so this restricts the single end read length to roughly 300 bases, and requires the paired end turn around to regenerate the template.
Secondly, the HiSeq uses fairly powerful lasers which also damages the DNA. The MiSeq on the other hand has LEDs, which are not as damaging.

[pmiguel]

My understanding that the read lengths are limited due to DNA damage.

Firstly, incorporation chemistry/removal of the dye terminator is somewhat corrosive to the DNA - so this restricts the single end read length to roughly 300 bases, and requires the paired end turn around to regenerate the template.

I doubt that this is the case. The limitation is more likely phasing/prephasing -- which has to do with the ensemble of product molecule containing members slightly behind or slightly ahead of the bulk of the molecules.
If the template molecules were actually damaged, then turn around would not be able to generate reverse-complement strands for sequencing the reverse read(s).

Secondly, the HiSeq uses fairly powerful lasers which also damages the DNA. The MiSeq on the other hand has LEDs, which are not as damaging.

Nope. The HiSeq 2500 in Rapid mode can do 2x250 base reads at roughly the same quality as the MiSeq does 2x250 base reads. So, whatever the issues are limiting the read lengths, it won't be the laser.
--
Phillip

[bbp]

Hmm. That was the answer given to me by Illumina. It's possible that they were not providing the correct information, but I'm inclined to believe it.

I see no increase in my phasing/prephasing rates on a HiSeq 2500 1x 200 cycle rapid run, and I'd have expected this to be cumulative. Perhaps someone with a 2x 250 rapid run could report their phasing/prephasing stats over the course of a run.

With regards to DNA damage, I'd suspect that the max cycle length has been optimised to ensure that information is not lost beyond some acceptable standard. If illumina could acceptably perform longer cycle reads, they most certainly would do it.

[pmiguel]

Wait bbp. You are sure this is really Illumina who gave you that answer? Because Illumina usually sounds just as portrayed by GenoMax above.
And if it isn't Illumina, then who is it?
I didn't mean there was a difference in phasing/prephasing between the HiSeq and the MiSeq. That isn't really an issue because they both produce basically the same read lengths max. But phasing/prephasing is ultimately the main limiter of Illumina read lengths.
Also, the marginal improvement in useful results when moving from 150 base to 250 base reads is small for most users. The usefulness of the 250 base reads tends to be in edge cases. So the desire to increase read lengths beyond 150 base on Illumina instruments will be low.
--
Phillip

[luc]

In my experience very few people at Illumina have an in depth knowledge of the sequencing chemistry.
I had also wondered about this question; why was the Miseq (especially in the good old days when the V3 chemistry came out) able to deliver high quality longer reads?
It seems to me that the HiSeq and MiSeq chemistries are optimized differetnly; the fluorescence levels tend to down with increasing cycle numbers for HiSeq runs and up for MiSeq runs.