Help to decide the optimal sequencing usage for 6 lanes

[vebaev]

Dear all,
In the next weeks a FAS of Illumina will came to install GAIIx and will do a training. It has been told us that he will do a test run with control DNA on the 2 lanes and we can use the other 6 lanes for our samples.
How to decide what is optimal usage for these 6 lanes. We have one unsequenced plant that we studied with genome about 1300MB what can we do - genomic DNA seq or to make several RNA-seq transcriptomics with stress conditions?

How to calculate optimal usage or when we choose RNA to DNA-seq we want to see how many samples to prepare?

thanks!

[vebaev]

Anyone ?...

[huguesparri]

Is the training based on a single read or paired end sequencing? What will be the length of the sequences?

[vebaev]

Single reads, we can choose the lenghts

[tokikake]

Hi vebaev,
we had our GAIIx training last week and did the following: We prepared three additional libraries, two bacterial WGS and one cDNA library. In our training there was no time for index sequencing, so we choose two dilutions (8pM and 10pM) for each library (including the PhiX) and load each on a separate lane to get a feeling what dilution works best for our samples.

By the way: Although we will do a lot of SR sequencing we did a 2x36bp paired end sequencing in the training to get familiar with the procedure and it was a good decision.

Tokikake

[vebaev]

Thanks, and is 1300mb genome is sutable for 1 lane how can we calculate it? Or it is better to put a transcriptom on the lane since it is smaller then the genome if there will be no indexing?

[tokikake]

Thanks, and is 1300mb genome is sutable for 1 lane how can we calculate it? Or it is better to put a transcriptom on the lane since it is smaller then the genome if there will be no indexing?

I am not sure if I get you right. Ho much is suitable depends on the size of your genome and the coverage you want and of course the sequencing experiment (output).
For a 2x36 bp sequencing experiment (or 1x72bp) on the GAIIx (with best 40 Mio reads per lane) you'll get 2,8 Gb per lane (we got 39 mio reads from the best lane in our training and it was not the control!). This would be enough for 19 bacterial genomes of 5Mb with a 30 fold coverage (for 454/illumina hybrid assemblies for example) if I calculated right.

But maybe there is an experienced user who can give you (and me) a better recommendation then a newbie like me

PS: we skipped indexing because otherwise we had to do the paired end procedure without the trainer (sequencing an index takes about 7 hours)

[vebaev]

thanks for the reply

[huguesparri]

In my humble opinion, for de novo sequencing of a rather large genome, I recommand you to use paired end sequencing with a maximum read length and try to have a 30x coverage.
In your case, using a paired end 2*100nt run and 1 lane, you will end up with:
2*100nt*30.000.000 reads (let's be pessimistic)= 6.000.000.000nt => a 3.3x coverage. If you use 6 lanes, you will end up with a 20x coverage. It might be enough (or not) depending on the compexity of your genome (repeated region, etc...).
If you use shorter reads (36nt for example), you will have hard time assembling them if you don't have a reference genome.

[vebaev]

Thanks @huguesparri and what about doing an RNA-seq for 6 conditions each on a separate lane? What you suggest SE PE lenght?