NGS Sequencing

[pavan]

Hello Everybody!!
Will the outputs of two different NGS machines (For ex : 454 and Illumina) remains same for the same input.

I say outputs never be same due to technical bias, inbuilt sequencing technology. Can any body tell what could be the other reasons?

And these different outputs, when fed to same assembler (for ex : Velvet)
Does the assembled data remains same? If no, why?

Thanks
Pa"1"

[pavan]

Metagenome NGS Reads

Hello Everybody!!

I just started my work in metagenomics.

Well i want to compare two metagenome datasets which are sampled in 2 different ways and also sequenced by 2 different NGS machines. (Data 1 with Pyrosequencing and Data 2 with Illumina).

How can i justify in comparing those 2 datasets (searching for pathogens), though they are sequenced by two different NGS technologies.

Thanks
Pa1

[krobison]

Hello Everybody!!
Will the outputs of two different NGS machines (For ex : 454 and Illumina) remains same for the same input.

I say outputs never be same due to technical bias, inbuilt sequencing technology. Can any body tell what could be the other reasons?

And these different outputs, when fed to same assembler (for ex : Velvet)
Does the assembled data remains same? If no, why?

Thanks
Pa"1"

Yes, for the reasons you mentioned. Plus, if you load the same sample into the same instrument, the data will not be exactly the same, as each draws a random sample of the input material.

For that reason, unless you have a very uniform sample, assembly results could be very different if you assembled from two different runs on the same instrument, given a highly diverse input DNA pool.

As far as assembly, Velvet is likely to do badly on 454 data -- assemblers tend to have a dependence on the error profile of the underlying data. Even with perfect data, the different structure of the data of 454 vs. Illumina (longer reads on 454; paired ends on Illumina) will give you different power to assemble a genome. Read length matters!!! For any given genome & perfect data, there will be regions that cannot be assembled using a particular read length (for paired-end, this is more complicated but is effectively the length of the insert).

[krobison]

Hello Everybody!!

I just started my work in metagenomics.

Well i want to compare two metagenome datasets which are sampled in 2 different ways and also sequenced by 2 different NGS machines. (Data 1 with Pyrosequencing and Data 2 with Illumina).

How can i justify in comparing those 2 datasets (searching for pathogens), though they are sequenced by two different NGS technologies.

Thanks
Pa1

First, are these shotgun metagenomes or 16S amplicon data? Which Illumina platform with which chemistry (2x100? 2x150? 2x250? etc). How deep is each dataset?

Yes, they are expected to be very, very different. You almost certainly sampled much, much more deeply with Illumina. Error profiles will be different on each one. Longer reads on 454 may be more revealing. Etc.