Horrors Of The Theoretical Maximum
Today is Halloween1 so I want to tell you a horror story. What if all vendors started quoting Maximum Theoretical Output like Oxford Nanopore! đ˛đ˛đ˛đ˛
PacBio
The Revio chip has 25 million wells. Theoretically each of those wells could be producing a read. That would be 375Gb/run (as opposed to the current 120Gb spec). But wait, thereâs more! Perhaps we should normalize this for a 72 hour run time to make it consistent with Oxford? And also, itâs not really reasonable to compare simplex versus PacBio Hifi⌠soâŚ
PacBioâs instruments incorporate about 3 bases per second. That gives a theoretical maximum of:
Oxford Nanopore
We donât need to guess in Oxfordâs case as they already state it as 290Gb. Wow⌠not even 1Tb!
Illumina
Illumina have a â25Bâ flow cell on the NovaSeq X. This produces 8Tb at 2x150bp. Assuming a PF rate of 80% weâre looking at ~10Tb. But thatâs only a 48h run, so letâs boost that by 1.5 giving us:
Summary
Of course, the above is just a horrible horrible dream of a world where vendors quote theoretical numbers, not the expected or guaranteed run output. One where you can never be sure how much data you will get from a run. And when faced with a poor run vendors can simply smile kindly and say:
âThe difference between theory and practice is greater in practice than it is in theory.â
Of course⌠this is not the world we live in. Or would want to live inâŚ
Itâs already the 31st here in Japan.