Differential amplicon size effects on clustering across Illumina instruments
 

I would like to recommend to your attention a Bioarchive (pre-print) paper:

https://doi.org/10.1101/388108

entitled:
Measuring Illumina Size Bias Using REcount: A Novel Method for Highly Accurate Quantification of Engineered Genetic Constructs

See panel D below:

https://www.biorxiv.org/content/bior...600&carousel=1

They add the same molar amount of a bunch off different size constructs and then measure the relative amount of clustering they get.

It looks like the NovaSeq is far less sensitive to amplicon size with regards to clustering efficiency! It only looks like ~20x drop in efficiency going from 150bp to 1500bp!! Whereas for a MiSeq it would be about 5000x drop in efficiency of clustering.

If you are a NovaSeq hater, feel free to hypothesize that this effect is largely the result of long amplicons being able to escape into adjacent (or even cross-surface) wells. Thus you can attribute the NovaSeq advantage to duplicate amplicon clusters...

--
Phillip

Interesting difference between NextSeq Rapid and HO. The rapid groups with the examp platforms, while HO looks like the other random cluster platforms.