More Protein Sequencing Thoughts
Portal has raised $35M for their nanopore protein sequencing platform, I wrote about them before, and I don’t think any new data has been announced with the funding.
With the new Portal raise, and my recent posts on other protein sequencing companies, I thought it would be interesting to briefly review.
Sequencing/Fingerprinting
Protein sequencing is more challenging than DNA sequencing in multiple different axes:
Single molecule, because we don’t have amplification approaches.
Much harder detection problem, looking at 20+ AAs rather than 4 nucleotides.
No simple way to incorporate labels as with sequencing-by-synthesis.
Much higher throughput requirements due to higher dynamic range.
Counting
On the positive side, protein sequencing applications are largely counting applications. That is, you will be seeing the same sequence hundreds of times. So, single read, single base fidelity may be less important.
This means that whatever you build, it probably ends up being more like a classifier than a single read sequence. It seems unlikely that you’d ever want fastq files filled with thousands of lines of the same sequence.