Illumina - naive questions

[yaximik]

Respected Illuminati,

I used to sequence with 454 GS Jr, now I am trying to figure out how MiSeq works. I read a b unch of manuals and user bulletins but I still cannot get a few things, so I would appreciate a little help from experts.

1. MiSeq is said can use only paired end flow cells, and does not support single end read framework, so paired end is the only option. Is that correct?

2. TruSeq DNA Sample prep LT is said can be used for preparation of single, paired end and multiplexed libraries. Yet it includes only single index adaptors, unlike the HT kit that includes double index adaptors. So how LT can be used for making paired end libraries?

3. It is said that a minimum two indices need to be used to ensure an adequate diversity. I presume it is meant two samples labeled with different indices need to be pooled. But what if I want to sequence just one mammalian-size genome, how can I do this this? Use the same library but with different indices?

4. The maximum output of MiSeq with 500 cycle kit is about 7.5 Gb, which is roughly 2x of human genome. GAIIx requires one $16K+ reagent kit and will run about 2 weeks for the maximum output. MiSeq requires a $900+ kit and will run about 30 hrs producing its maximum output. So, running MiSeq 14 times at cost of $13K in reagents will produce 7.5x14=105 Gb, which is comparable to GAIIx. So it can be used for human genome seqeuncing as well, but it is not recommended. Am I missing something here?

Thankls for help.

[huguesparri]

Hi,
I'm not a specialist about MiSeq as I only work on a HiSeq2000, but i think I can still answer some of your questions:
Question 2: I can tell you that indexing a library and sequencing it using SR or PE protocol is compatible. With the Truseq adapters, you can perform SR (Single Read), PE (Paired End), SR indexed and PE indexed runs at your will. On the hiseq2000, index sequencing is independant from your DNA of interest sequencing. I suppose it's the same on the Miseq.

Question 3: If you only have one library to sequence, you will not have to perform an indexed run. On the Hiseq 2000, it's just an option you can choose (or not) when you are parametring your run. I suppose it's more or less the same on the Miseq.

[microgirl123]

We have a very new MiSeq so I'm not an expert (and we only have done bacterial genomes so far, but if you only have one library on the flow cell, there is no need to index at all. If you do use an indexed adapter, you can just ignore it (according to Illumina) and not do the index read. I have done this and had no problems.

[yaximik]

Thanks, guys. Just to make sure I understood - one can use adaptors with whatever indices for making a library and just specify non-multiplexing (no index read) workflow in experiment setup, correct?

I am still confused about paired end workflow, which in my understanding requires reading from both ends labeled with different indices... Any help, please?

[Klaus Metzeler]

I guess the key point here is that the index reads and the main sequencing reads are separate reactions (they use different priming sites in the Illumina TruSeq adapters). A typical run setup may be: 151bp forward read - 8bp Index read 1(from F adapter) - 8bp Index read2 (from R adapter)- 151bp reverse read. These are four entirely separate and independent reads.
If you have a single sample, you can simply omit the index reads, and just run 2x151bp sequence reads. IF you do index reads, however, you need to make sure that you have a sufficient degree of index diversity.
You can also do a paired-end run, but choose to only read one index. There is no technical necessity to read both indices during a paired-end run.

If you want, you can also set up the sample sheet so that the MiSeq performs a single-end 300bp read with a 300bp reagent cartridge (although, of course, your sequencing quality will degrade towards the end...). You have to use Illumina paired-end adapter sequences because only those will anneal to the flow cell surface and work during cluster generation. However, that doesn't mean that you are actually forced to do a paired-end sequencing run.

[kmcarr]

1. MiSeq is said can use only paired end flow cells, and does not support single end read framework, so paired end is the only option. Is that correct?

Even though the flow cell is capable of producing paired end reads you can configure the run to produce single end reads up to 500bp if that is what you desire.

2. TruSeq DNA Sample prep LT is said can be used for preparation of single, paired end and multiplexed libraries. Yet it includes only single index adaptors, unlike the HT kit that includes double index adaptors. So how LT can be used for making paired end libraries?

The number of indexes used to identify which library a read comes from is completely independent of the number of sequencing reads performed.

3. It is said that a minimum two indices need to be used to ensure an adequate diversity. I presume it is meant two samples labeled with different indices need to be pooled. But what if I want to sequence just one mammalian-size genome, how can I do this this? Use the same library but with different indices?

If you are sequencing a single library an index read is unnecessary so you wouldn't include one.

4. The maximum output of MiSeq with 500 cycle kit is about 7.5 Gb, which is roughly 2x of human genome. GAIIx requires one $16K+ reagent kit and will run about 2 weeks for the maximum output. MiSeq requires a $900+ kit and will run about 30 hrs producing its maximum output. So, running MiSeq 14 times at cost of $13K in reagents will produce 7.5x14=105 Gb, which is comparable to GAIIx. So it can be used for human genome seqeuncing as well, but it is not recommended. Am I missing something here?

If you are doing Human genome sequencing then neither the MiSeq or GA IIx are the right platform. You should be using the HiSeq. One lane of 2x100bp PE reads on the HiSeq will generate ~30Gbp (10x Human genome) of data and there are 8 lanes available on each HiSeq flow cell.

[kerryp]

running just one library on MiSeq (Nextera XT)

Just wondering how to go about running a single bacterial library (rather than multiple pooled bacterial libraries) on MiSeq. So far I'm only going from a post in a thread from a few years back, quoted below:

We have a very new MiSeq so I'm not an expert (and we only have done bacterial genomes so far, but if you only have one library on the flow cell, there is no need to index at all. If you do use an indexed adapter, you can just ignore it (according to Illumina) and not do the index read. I have done this and had no problems.

How does this work? Do I have to do anything differently in the library preparation, or is it just a matter of telling the Miseq not to read the indexes? I assume the index primers still have to be used in order to amplify the product. Where in the Miseq run setup do I tell it to ignore the indexes?

Thanks in advance for any advice!

[microgirl123]

When you set up your sample sheet in IEM just select 0 indices. Then the MiSeq won't do the index read. Keep in mind that you cannot add phiX to a single library since it won't be able to sort the unindexed phiX out from your library.

[kmcarr]

How does this work? Do I have to do anything differently in the library preparation, or is it just a matter of telling the Miseq not to read the indexes? I assume the index primers still have to be used in order to amplify the product. Where in the Miseq run setup do I tell it to ignore the indexes?

There is no difference in library prep; you will still be using the Nextera XT index primers, but just one pair of them (i.e. one N7xx and one N5xx). The user does not directly configure the Index reads on the MiSeq, rather the MiSeq Control Software determines the index read configuration based on information you have entered in the SampleSheet. When you create your SampleSheet (e.g. using Illumina Experiment Manager) simply select "0" for number of index reads. The MiSeq will then simply skip any index reads. All reads from the run will be output into a single pair (R1 & R2) files.

It is still perfectly acceptable to enter the index information in sample sheet even with a single library. In this case the instrument will perform the required index reads and "demultiplex" your data at the end. If you are spiking in PhiX control library this is a convenient way to separate your library reads from PhiX reads.

[Cross-posted with microgirl123]

[microgirl123]

kmcarr - doesn't the MiSeq have trouble with the index reads if it's only seeing one index?

[kmcarr]

kmcarr - doesn't the MiSeq have trouble with the index reads if it's only seeing one index?

Generally no, at least not with the latest iterations of Real Time Analysis. If it so happens that the index read is unusable one can always go back and reprocess the run data ignoring the index and outputting everything to a single file. Nothing is lost in this case.

[westerman]

Generally no, at least not with the latest iterations of Real Time Analysis. If it so happens that the index read is unusable one can always go back and reprocess the run data ignoring the index and outputting everything to a single file. Nothing is lost in this case.

And going along with this, if you do ignore the index then you can always bioinformatically remove the phiX. I happen to be doing this at the moment with a couple of projects. One project had about 75% reads demultiplexed and I felt that throwing away that many reads (I should have gotten closer to 95% of the reads with my single index) was not acceptable.

[kmcarr]

Thought I should post an example of a single library MiSeq run with index:

- One TruSeq Small RNA library (Index RPI22) + ~0.5% PhiX

- MiSeq v2 Standard flow cell, 50 cycle reagent kit

- MCS v2.5.0.5/RTA v1.18.54

- Cluster Density 784 K/mm^2 (15.1M), 94.7% PF (14.3M)

- 98.5% of reads assigned to expected index and of those 99% were perfect match.

Below are the Mean Q30 and % Base plots.

[kerryp]

Many (belated) thanks!