Semyon Kruglyak, PhD - VP of Informatics at Element Biosciences
- AVITI runs routinely produce a high fraction of Q40 bases
- The predicted Q scores are accurate
- Third-party, open source tools are used to assess quality score accuracy
- FASTQ files supporting the Q40+ claims in this talk are freely available for download here
Transcript of Q&A Session
Q: How hard is it to hit the density target when deciding on library concentration?
A: We have a quantification step that helps. You may see some variability when first trying, but then you dial it in. We also accept a fairly high range for the density.
Q: Is it a patterned flow cell?
A: It is a random flow cell.
Q: What are the read lengths?
A: Our data is generated from 2x150 for this talk. You can of course go shorter.
Q: What types of libraries have you run?
A: We have a long list of library types through a broad ecosystem of partners — whole human genome, RNA, single cell, exomes, panels, and more.
Q: What do you see as your position in the market?
A: We want to lead with a quality, while offering a cost-effective benchtop solution. You can do runs in your own lab with full flexibility and you don’t have to send out to get the economies of scale.
Q: What is the price of the instrument?
A: The sequencer is $289,000 and a 2x150 kit is $1,680 (sequencing reagents and flowcell).
Q: Is the current data filtered?
A: There is minimal filtering. If a feature has low quality in early cycles, we filter it. In general, we filter 5% of our data or less, and the rest goes into the FASTQ file that you download. There’s no other manipulation beyond those early cycles of R1 and R2, with the exception of optional adapter trimming.
Q: Are the errors random?
A: The errors seem mostly random. Since we are fairly early in our review it’s something that we’re still looking into.
Q: How do you handle low diversity?
A: We require high diversity in the first few cycles of Read-1 so we can find where all the features are. Beyond that, we have a quite robust low diversity pipeline that handles the type of challenges that low diversity presents.
Q: Have you sequenced bisulfite converted DNA (i.e. methylseq)? And does data from methylseq have great quality scores as well?
A: Yes, we have, and they seem to work pretty well, though density is somewhat reduced. We would love to partner with someone to optimize performance on these libraries.
Q: How long does it take to run the 2x150?
A: Just under 48 hours.
Q: Do you support dual indices?
A: Yes, we support dual indices and support two types of library prep. If you have your own linear library, we will take it and apply our compatibility kit and then run it with however many indices you have. If you use our Elevate library, then we have unique dual indices for 96 plex and many more if you do not need UDI.
Q: Are dual indices needed on this platform?
A: Although our amplification methodology does not appear to lead to index hopping, they may be considered for other reasons. For example, if your index plate is all synthesized together, you can have oligo contamination — leading to index misassignment. This is something that we’re currently exploring.
Q: What’s the duplicate rate?
A: We are seeing very low duplicate rates. Typically sub 1%.