Taylor priest
@taylorpriest.bsky.social
NOMIS-ETH fellow at Institute of Microbiology and Center of Origin and Prevalence of Life, ETH Zurich | microbial ecologist interested in diversification, evolution and mobile genetic elements | Parent of a beautiful Briard | food and fitness enthusiast
And a very worthwhile mention ;) globDB is an extensive database so you would get even better coverage than using GTDB as a reference !
January 29, 2026 at 11:45 AM
And a very worthwhile mention ;) globDB is an extensive database so you would get even better coverage than using GTDB as a reference !
Good point. Of course this approach would exclude reads without genes. I have used this approach for PacBio HiFi reads before and it works well, but there you rarely get any reads shorter than 1 kbp. I imagine with Nanopore this is only a minor fraction of the reads though, right?
January 29, 2026 at 7:06 AM
Good point. Of course this approach would exclude reads without genes. I have used this approach for PacBio HiFi reads before and it works well, but there you rarely get any reads shorter than 1 kbp. I imagine with Nanopore this is only a minor fraction of the reads though, right?
True, it could definitely be used to get compositional insights. I am not convinced by the accuracy of the relative proportions from it though.
January 29, 2026 at 7:03 AM
True, it could definitely be used to get compositional insights. I am not convinced by the accuracy of the relative proportions from it though.
Could also do whole read alignments using minimap2 to e.g. GTDB genomes and the LCA if multiple hits found
January 28, 2026 at 7:15 PM
Could also do whole read alignments using minimap2 to e.g. GTDB genomes and the LCA if multiple hits found
I don't believe there is a 'best practice'. In the past I've used a protein+LCA based approach. LCA helps to give taxonomic insights even when there aren't high identity matches. Can use GTDB or any other genomes database as a ref (have to build a non-redundant protein catalogue from it)
January 28, 2026 at 7:12 PM
I don't believe there is a 'best practice'. In the past I've used a protein+LCA based approach. LCA helps to give taxonomic insights even when there aren't high identity matches. Can use GTDB or any other genomes database as a ref (have to build a non-redundant protein catalogue from it)
Great paper! Well done to everyone involved 👏
January 23, 2026 at 10:38 AM
Great paper! Well done to everyone involved 👏
Reposted by Taylor priest
Don’t forget that the 1961 triplet code paper only became elegant when it was written! The experiments were all a bit of a mess... I suspect the same was true of the other classic papers!
A breakthrough from 60 years ago: “General nature of the genetic code for proteins” (1961)
In 1961, Francis Crick and Sydney Brenner, together with two Cambridge colleagues, published an article in Nature that used simple genetic experiments to demonstrate that the genetic code was almost ...
onlinelibrary.wiley.com
January 11, 2026 at 3:47 PM
Don’t forget that the 1961 triplet code paper only became elegant when it was written! The experiments were all a bit of a mess... I suspect the same was true of the other classic papers!