What does it mean by "150/350 PE reads" from MiSeq?

[ymc]

http://www.ncbi.nlm.nih.gov/sra/SRX573008[accn]

I downloaded this SRA data. I noticed that the R1 reads are 150bp but the R2 reads are 350bp. How could that be? Aren't they supposed to be the same length???

[nucacidhunter]

Unequal length on MiSeq is permitted.

[ymc]

Unequal length on MiSeq is permitted.

So this data was generated by a 2x250 run?

Why do people want to do unequal length?

Thanks in advance!

[Brian Bushnell]

All else equal... longer reads are more useful. 350bp reads can span repeats that 250bp reads can't, for example.

However, all else is NOT equal, and I seriously doubt the quality at the end of those 350bp reads is very good.

[ymc]

All else equal... longer reads are more useful. 350bp reads can span repeats that 250bp reads can't, for example.

However, all else is NOT equal, and I seriously doubt the quality at the end of those 350bp reads is very good.

Thanks a lot for your reply. Are there any limitations to adjusting the lengths? Can you do 1/499 or even 0/500???

[HESmith]

Well, you need 4-5 cycles in the first read to call clusters, and at least 25 cycles for base-calling. Plus, as Brian noted, the quality drops precipitously with the longer reads. But I think that the instrument could be programmed for 25/475.

[ymc]

Well, you need 4-5 cycles in the first read to call clusters, and at least 25 cycles for base-calling. Plus, as Brian noted, the quality drops precipitously with the longer reads. But I think that the instrument could be programmed for 25/475.

Thanks for your reply. It is good to know that this is possible anyway. 25/575 for the 2x300 kit can come in handy sometimes.

[GenoMax]

Or you could do a 600 cycle SE run.

[ymc]

Or you could do a 600 cycle SE run.

So you can turn PE into SE in V3 chemistry but not V2?

[nucacidhunter]

It can be done with v2 too and possibly on HiSeq. In MiSeq after cycle 200 the quality drops and is in lowest value at 300 cycle. Increasing cycle beyond that results in low quality so it may not be useful depending on intended use of data.

[matth431]

So this data was generated by a 2x250 run?

Why do people want to do unequal length?

Thanks in advance!

One reason we've done this is where there's a high proportion of short fragments in the library - we've had a user make Nextera libraries from very low amounts of viral DNA which tend to over-tagment, so lots of ~200bp inserts. If the bulk of the reads sequence past the end of the molecules, the machine loses focus and the run will abort prior to the index reads.

So we often do 100/400 reads to ensure we get index data for all clusters, then hopefully get as much of a reverse read as possible for the small percentage of larger fragments in the library - if the run aborts on read 2, we can at least recover the data, but if it aborts on read 1, the data is useless as it can't be demultiplexed.