Illumina HiSeq library insert size

[SDPA_Pet]

2X150bp only can sample 300 bp from a fragment (for what ever size fragment, as long as it can get sequenced).

Choice of insert sizes depends on what you are trying to do. If you have a reference available then making the libraries so the two ends do not overlap makes sense since you can sample a larger region. If you must have the entire region covered by the two reads (i.e. reads need to overlap) then you would want to make inserts smaller.

Which of these two cases were you wanting to do?

I do not have reference genome. My samples are environmental samples from soils. As I said, I got quit good joined ratio > 50%, which surprised me, because the reports told me the average insert size is 600 bp.

[GenoMax]

Hi Genomax,

Thanks. What you said makes me think the sequencing center send me a wrong report. They might mean the largest fragment. It doesn't make any sense for them to build so large fragment. 2X150bp only can sequence 300 bp maximum. If they build a library size of 600 bp, there are 300bp gaps out there. The coverage won't be very good.

Thanks,

2X150bp only can sample 300 bp from a fragment (for what ever size fragment, as long as it can get sequenced). You also need to keep in mind that there will always be a "normal" distribution of fragment sizes in your library with some tailing on both sides. How those tails look may determine the outcome of what preferentially clusters (small fragments would) on the flowcell.

Choice of insert sizes depends on what you are trying to do. If you have a reference available then making the libraries so the two ends do not overlap makes sense since you can sample a larger region. If you must have the entire region covered by the two reads (i.e. reads need to overlap) then you would want to make inserts smaller.

Which of these two cases were you wanting to do?

[SDPA_Pet]

Hi Genomax,

Thanks. What you said makes me think the sequencing center send me a wrong report. They might mean the largest fragment. It doesn't make any sense for them to build so large fragment. 2X150bp only can sequence 300 bp maximum. If they build a library size of 600 bp, there are 300bp gaps out there. The coverage won't be very good.

Thanks,

[GenoMax]

BBMap is going to give you an absolute answer by actually using the data that is there. There is no ambiguity involved. It will work if you have a reference available or without. Only case it won't work is if you have reads that don't merge and you don't have a reference available.

If you are able to join the PE reads then there are some inserts there that are smaller than 300 bp.

While you library may have had fragments in the 600 bp range, if there were any that were of a smaller size (as indicated by tails on bioanalyzer traces, you don't get an an absolute answer from bioanalyzer, AFAIK) then those fragments will preferentially bind and form clusters.

[SDPA_Pet]

Hi GenoMax,

Yes, I know the bioinformatic tools BBMAP. According to their report, it says they determine the size of library using Agilent 2100 Bioanalyzer. I have never used a Bioanalyzer. I would guess it is kind of instrument that can do physical measurement (not a bioinformatic tool). Do you suggest that their reports or measurements are wrong. I should use bioinformatic tools to check it? Is it common that bioanalyzer gives you a wrong number?

So, I am correct, right? To join 2X150 bp, most of inserts should be less than 300bp.

[GenoMax]

Use BBMap to estimate insert sizes. There are two methods described here. That estimate of 600 bp is clearly wrong since you would not have been able to merge the R1/R2 reads otherwise.

2x250 is maximum supported length on HiSeq 2500 and 2 x 300 on MiSeq. One can't get longer sequencing lengths on currently available Illumina sequencing kits. One could run asymmetric runs (e.g. 1 x 600 bp) but that is not generally recommended due to drops in quality you are bound to experience towards the end of such runs.

[SDPA_Pet]

Hello, I did a illumina HiSeq 2X150 bp metagenomic sequencing recently. I have some questions.

1>I have got my sequencing report back (from sequencing center). The report says the average insert size is about 600bp, which means majority of reads that was prepare to be sequenced are around 600bp. I am confused about it. You know, after I got my fastq files back (R1 and R2). I firstly merged paired ends. I have >60% of reads that can be join together successfully. I don't how could this happen. Since the method only sequence 150 bp, and the fragment is 600 bp. There will be no overlaps (150 X 2 = 300 bp << 600 bp). Why I can still get so many reads joined. Let says, if I want to join more paired - end reads, the fragment size should be designed less than 300 bp right?

2> The report also says "300 cycles using the HiSeq system". This straight-forward. I suppose for R1 and R2 is 150 cycles, receptively. Each cycle will add one nucleotide and 150 cycle will be 150 bp. The sequencing center says they can also do maximum 500 cycles, which means 2X250 bp sequencing. I was wondering why they don't run more cycles such as 1000 cycles, so we could get 2X500 bp. This will give us longer reads. I don't know which factors restrict the illumina reads lengths? For the reports, it seems we can increase cycles to get longer reads.

Thanks,