PacBio RS II SMRTcell performance variability

[ribodoc]

We are observing increased variability in the performance of SMRTcells by lot, strip and within a strip. Using the same SMRTbell template on different sequencing runs the variability can be as much as a two fold difference between strips of SMRTcells. We would like advice from other users who may experiencing the same issue.

[ECO]

Are you referring to just successful loading? Or some other metrics (RL, accuracy)?

[GenoMax]

@ribodoc: Was there a time when you were regularly getting consistent results and now you are not?

[ribodoc]

Are you referring to just successful loading? Or some other metrics (RL, accuracy)?

Within a week, with the same SMRTbell template and loading concentration went from 47.44% P1 to 26.64% P1. Accuracy dipped from 0.841 to 0.831. Polymerase read length N50 had a smaller drop from 16285 to 15849.

[ribodoc]

@ribodoc: Was there a time when you were regularly getting consistent results and now you are not?

Occasionally we see this occur and it was time to put it out to the community. We typically do a titration of SMRTbell concentration for loading and then do a production run on a strip of SMRTcells. There will be periods of time when the tritration and production runs are very consistent and then there are other times when they are not consistent and even not consistent within a strip of SMRTcells.

[GenoMax]

Occasionally we see this occur and it was time to put it out to the community. We typically do a titration of SMRTbell concentration for loading and then do a production run on a strip of SMRTcells. There will be periods of time when the tritration and production runs are very consistent and then there are other times when they are not consistent and even not consistent within a strip of SMRTcells.

That has been more or less the story. While I don't work at the bench I hear similar observations most of the time. There is no logical explanation that can be put forth either. Hopefully Sequel will be more reliable.

[ribodoc]

That has been more or less the story. While I don't work at the bench I hear similar observations most of the time. There is no logical explanation that can be put forth either. Hopefully Sequel will be more reliable.

Thank you for the response.

[ECO]

@Genomax...the logical explanation is that the SMRTcell production process is dramatically more complicated than all other sequencing consumables, ZMW size and uniformity, the micromirror shape, aluminum coating, then on top of that the blocking of the surface and then the biotinylation.

I know it's no solace to scientists trying to do reproducible production (we run two RSII's all the time), but it's kind of amazing just how well it all works, and is a testament to all the people that have contributed at PACB. (disclaimer I worked there for a long time and know a lot of people there still)

[GenoMax]

I have nothing but admiration for the amount of technology that went into making the original PacBio machine, consumables.

Based on anecdotal references keeping an RSII running ALL the time seems to be essential for its well being. Sadly we could not do that.

[luc]

The SMRT-cell quality (and correlated to it the yields) can indeed be quite variable in general .
The very last batches of SMRT-cells seem to be more variable than usual (for us and also some other labs). These higher frequency problems seem to occur at least once a year.