Is The Revio A Single Color System?

Many years ago I wrote about QuantumSi and compared their chips to PacBio’s. At that time it seemed relatively clear that the PacBio design used 2 channels. Their patents show a stack up with two filters built into the chip, and a SEM image of the same device:

This some ways somewhat similar to Illumina’s 2-color sequencing approach, two filters/sensors are used to read out 4 colors. The difference here being that there’s no “dark” base¹. Different fluorophores just gives different patterns of intensity across two spectral bands:

Figure 3 shows Howa readout over a small spectral band can be used to distinguish between two fluorophores. Figure 4 shows this in a 4 color context, using two sets of filters.

But recently I was prompted to go hunting again to see if there had been any new developments and came across this patent. The patent shows what clearly appears to be a single channel system and provides a justification for this:

The patent describes “devices with reduced-scale unit cell”. Which one might exactly is exactly what would be required to enabled a throughput jump from the Sequel II (8M) to the Revio (25M) without a massively larger chip.

That is to say a single channel sensor using 4 different intensities at a particular wavelength². This would halve the number of sensors required for the same throughput as well as simplifying the fabrication process. The question is… are they doing this or not?

One clue perhaps is the SNRs (which are encoded in the BAM files³). SNR differs significantly between nucelotides:

PacBio SNRs by nucleotide extracted from a Revio BAM file (here from the Pangenome project). X-axis is also changing, but A: ~11, T: ~7 G: ~4.5 C:~16

It’s interesting that the SNR varies so much and suggests to me differing brightness and a brightness order from brightest to dimmest of: C,A,T then G.

Where things fall down is that there’s no experimental support/SEM images in the single color patent. And all this could be consistent with a 2 color system, perhaps with SNR estimated derived from the combined signal.

So, I’m still pondering where to find additional support for the Revio being either a 1 or 2 color system. If you have any suggestions on where to look, please reach out (new@sgenomics.org or on the Discord!)

1. A dark base wouldn’t work in a non-cyclic system such as that used by PacBio.

2. or more likely range of wavelengths.

3. m84039_230117_233243_s1.hifi_reads.default.bam

samtools view m84039_230117_233243_s1.hifi_reads.default.bam | awk '{for(i=1; i<=NF; i++) {if ($i ~ /^sn/) print $i}}' | awk 'BEGIN{FS=","}{print $4}' > snrG