I think. This seems like good news from PacBio (boldface mine):
With regard to sequencing runs, as was the case in the previous three years, we expect to deliver another ~4-fold increase in throughput, reaching >4 Gb of data per SMRT Cell run, with average read lengths increasing to 15-20 kb. We plan to accomplish this through a combination of improvements in the sequencing chemistry, protocol workflows, and software. An example is active loading to increase the efficiency of loading one polymerase per ZMW at frequencies greater than the Poisson limit. In the area of data analysis, we will continue to work with the bioinformatics community to create faster algorithms for de novo genome assemblies, further developing solutions like our FALCON assembler for resolving diploid or polyploid genomes, and streamlined analysis workflows for other applications such as Iso-Seq™, full-length HLA, and others.
It is exciting to think about the new frontiers in genomics research that will be realized by this continued innovation and performance increases in SMRT Sequencing, for example:
• High-quality population and disease-specific human reference genomes
• Comprehensive views of tissue and disease-specific transcriptome architectures
• High-quality plant and animal reference genomes and transcriptomes
• Comprehensive characterization of structural variation in genomes
• Large-scale microbial genome and epigenome studies
If we assume you need around 500 Mb per bacterial genome (give or take), we’re talking about seven or eight bacterial genomes per run. That’s actually affordable in terms of cost-per-plate, and the turn around time (with some robotics) of sequencing one hundred bacteria is comparable to Illumina Hi-Seq. While the machines still take up a lot of space and are far more expensive than an Illumina Hi-Seq, this could be pretty big for microbial genomics, especially in high-end clinical labs.