NGS assembly

[pavan]

Hello everybody..

I have few questions about NGS reads.

1. Lets say 2 data sets mRNA enriched.
First data is sequenced by 454 and Second data is sequenced by some other
sequencer. And those outputs are assembled by some assembler.
Can i compare both he outputs (post assembly) by doing sequence analysis.

Putting it in simple way...
Can i compare the outputs of two different NGS machines. If not, why?

[sphil]

Sure you can, the problem is that each NGS machine has it's own field of application. While 454 or the newer version with longer reads pacbio etc. are commonly used for assemblies of genomes. Illumina machines are widely used for RNA-seq and stuff which does not "need" that long sequences. Though, you can also create assemblies from Illumina machines and compare them but the quality (of the assembly) might be different. This in fact, might have changed since Illumina kind of upgraded their machine to version 2.5 where you can sequence 2x150bp which is, counting an insert size of say 400-600 bp longer than the standard 454 library prep. Hence, imho there are only a few guys out their still using 454 in the future but thats another story...

TL;DR: You can compare the assembly outputs!

[pavan]

Thanks Sphil,

Can i compare two datasets out of which one is assembled and other is unassembled data?

Thanks
Pa1

[GenoMax]

Thanks Sphil,

Can i compare two datasets out of which one is assembled and other is unassembled data?

Thanks
Pa1

What kind of "comparison" are you planning to do?

[pavan]

Lets say Sequence analysis for finding out the antibiotic resistance genes in the datasets.

For Example:

Case 1 : Running BLAST : Assembled data(Input Dataset 1) Vs ARDB database and Raw data (Input Dataset 2) Vs ARDB database.

Now i have assembled the dataset 2
Case 2 : Running BLAST : Assembled data(Input Dataset 1) Vs ARDB database and Assemble data (Input Dataset 2) Vs ARDB database.

Now in both cases, Blast results remain same??
If the results are not same...What could be the exact reason.

Thanks
Pa1

[GenoMax]

Assembly is allowing you to reconstruct the "original" sequence (which was sampled by the reads you obtained). As long as the assembly is reconstructing the original sequence correctly (that is something you have to decide by looking at the data) you should get comparable results from the two blast searches.

Results from the "assembled" data would likely be more precise since you will have longer sequences to search with as opposed to raw data (which may show multiple hits if there are any repeat sequences).

[pavan]

Thanks a lot...
If I use 2 different assemblers on same input data, will the outputs of 2 assemblers look alike?

[rhinoceros]

Thanks a lot...
If I use 2 different assemblers on same input data, will the outputs of 2 assemblers look alike?

Depends on the assemblers, your data, and the settings you choose.

[sphil]

Thanks a lot...
If I use 2 different assemblers on same input data, will the outputs of 2 assemblers look alike?

I highly doubt that you will get the same output. Even two guys using the same assembler will not come up with the same results. (see Assemblathon resumee)