Results for the v2 PrecisionFDA Competition are out ( https://precision.fda.gov/challenges/10/view/results).

A thread. First, these highlight huge leaps in sequencing and informatics made possible in large part by the continuing work of NIST and @GenomeInaBottle. 1/

On the instrument side, the improvements by @PacBio (which has the highest accuracy in each method and category) and @Nanopore (able to deliver competitive SNP accuracy with Illumina across the genome) are clear. Not only is future of these bright, the present is as well. 2/

On the informatics side, the many high quality submissions from multiple groups highlight the health, robustness, and diversity of development in the field. Our team was happy to win highest overall accuracy with PacBio HiFi data and tie for this on multi-technology. 3/

In a collaboration with @kishwarshafin @benedictpaten @ucscgenomics we were happy to win highest overall accuracy with Oxford Nanopore data. Available here: https://github.com/kishwarshafin/pepper/blob/master/docs/PEPPER_variant_calling.md 4/

An approach by @edicogenome to remap difficult regions and use graphs from long read population sequencing data (if I understand the method) seems very strong. I look forward to getting my hands on this approach and seeing if we can complement it with other (and our) methods. 5/

On the MHC side, it seems that a number of teams have innovative approaches, particularly some graph-based methods that look strong, with @sevenbridges taking an award for MHC accuracy and @sentieon some others. This competition shows promise in graph and pan-reference methods 6/

I am impressed with the broad performance of @sentieon and @m_sahraeian, who submitted generally high performing results for all technology categories and each won a number. I am sorry I haven’t deeply delved into all the other submissions, I am sure there are more innovations 7/

A thank you to @PrecisionFDA and @GenomeInaBottle for making this competition possible.
The amount of advancement that this work drives is very substantial. Thank you to all of the participants, I look forward to learning from your insights and building on your work. 8/8