2. The reading length is 28bp. 50x coverage, a WGS, 10000 WGS per year, if a run runs for 8 days, the throughput of a run is about 32.8T in terms of data volume, which is definitely # of the size of an aircraft carrier in the sequencer, and of course, the floor space is also #. According to the official statement, the accuracy should be "unbelievable accuracy", and the error rate is about 1E-6, which is obviously far higher than the correct rate of any interviewed sequencer. However, please don't ignore that the reading length of CG sequencer is only 28bp, but please consider cutting the reading length of illumina from 250bp and 150bp to 28bp, which is more accurate.
3. 1 advantages first. The data output is huge, so the unit cost must be very low, which is suitable for human genome resequencing. CG has been advertised as "the best sequencing platform for human genome resequencing" since its birth. The "human" genome (or exon) with a reading length of 28bp and low complexity is very easy to handle. 28bp can only be re-sequenced, so let's sleep and assemble something. For NIPT and similar technologies, the sequencing process is just "accumulating data", and only the change of copy number needs to be detected, without paying attention to the change of coverage and SNP level. The new machine CG may be a very good choice (I'm not sure whether the length of 28bp may cause some troubles for large-scale mixed samples, and many bar codes with high complexity are needed). Similar to the detection of free tumor cells (ctC), free tumor DNA(ctDNA) and fetal nucleated red blood cells (FNRBC) for early noninvasive screening of tumors in the future, most of the measured data will be useless, and the advantages of CG as a "data free" platform may appear. It's just a bunch of data We're good at this.
3.2 Disadvantages, too much data is also a disadvantage-if you are not satisfied with running, most people can't run at all. According to the embarrassing situation of hiseq X, reading dragons is a serious injury, so there is almost no way to apply it to other fields except "human" and "re-sequencing" (far from the panda with complex human genome, probably ok). And the influence of 28bp on the bar code in pooling. Others refer to the machines before CG, whose running stability is a test. After all, 8 days is not short, and the amount of response data is huge. I wonder if it will be easy to need "after-sales".