Question: About possible bias due to different experimental design
0
gravatar for biotech
3.3 years ago by
biotech540
United States
biotech540 wrote:

Do you think I could have bias due to different DNA extraction kit, library preparation and Illumina version used?

The idea is to perform genome comparison of these different bacterial strains we have sequenced.

bias experimental design • 1.2k views
ADD COMMENTlink modified 3.3 years ago by chen1.9k • written 3.3 years ago by biotech540
2
gravatar for Brian Bushnell
3.3 years ago by
Walnut Creek, USA
Brian Bushnell16k wrote:

Yes. All kits have bias. All Illumina versions of anything (platform, software, etc) have bias too. So your data absolutely has bias, which is one reason it's best to only compare data using identical methodologies. If, for example, you wanted to do some kind of quantification with group A using kit W and sequencer X, and group B using kit Y and sequencer Z, your results would be complete garbage.

If all you care about is genome assembly, and not quantification, then it typically does not matter at all as long as you have sufficient coverage.

ADD COMMENTlink modified 3.3 years ago • written 3.3 years ago by Brian Bushnell16k

@Brian: Not sure if the sequencers will have bias (long as the same chemistry is used). I recall Illumina asserting that irrespective of machine, chemistry used the sequence one gets should be equivalent (that was before the advent of 2-color chemistry). The Q-scores may vary depending on chemistry/machine but the basecalls themselves should remain the same. If that had not been true then we would not have been able to use these technologies reliably.

ADD REPLYlink written 3.3 years ago by genomax69k

Thanks for your answer @Brian. The idea of the project is to perform genome comparison between 20-30 bacterial strains, of course preceded by genome assembly. Some of them have already been sequenced. One with PacBio and the other 4 with illumina. Do you think I should convince my PI to trow all they work away and sequence all 20-30 strains using identical procedures?

ADD REPLYlink modified 3.3 years ago • written 3.3 years ago by biotech540

Well, let me clarify my statement. It would only be a major problem if you are doing quantification, for example, trying to determine the relative abundance in a metagenome. There's no problem with assembly. Of course, you'll get different qualities of assemblies depending on your input data, and PacBio will better assemble the repetitive parts of the genome, but that's not an issue if you want to compare genomic features. It sounds like you are sequencing these as isolates, in which case you wouldn't be able to do quantification anyway.

ADD REPLYlink modified 3.3 years ago • written 3.3 years ago by Brian Bushnell16k
0
gravatar for chen
3.3 years ago by
chen1.9k
OpenGene
chen1.9k wrote:

Bias happens everywhere, especially for different platforms, different kits or different panels.

Remember to normalize your data before you do further analysis.

ADD COMMENTlink written 3.3 years ago by chen1.9k
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 878 users visited in the last hour