Greetings, no link or "comprehensive video" but you seem to miss denovo, ie. "from scratch" sequencing as it occured with covid and many others. To reuse your analogy, my understanding of an Alumina mechanical sequencer is that rather than build the portrait made of hundreds of smaller portraits based on an algorithm, imagine the image, actually a bunch of this original image, already existed but has been snipped in to bits of a random size a little bigger than the "small portraits". By doing lots of sequencing, or what would be grabbing the cut up bits and reading the pixels that make them up, even with errors, you can sort of put the original image back together based on the mathematical process of overlap. Some bits will contain overlapping segments because the bits don't come from just ONE image of the full portrait, but from a bunch of those portraits all cut up in to random bits. Its the overlap which allows a machine to read in and then analyze the most likely original, even considering an error rate in reading in the chopped-bits. This also means you can have other sequences of some stuff mixed in and still sort it out, at least that is my understanding of the basic mathematics. This also means you don't have to fully "isolate" since all you have to have is bits of DNA of sufficient length, read them in quick without too many errors, and rest can be done in software using math.