"Strains" section
Little by little, I'm finishing up the website. It now has a Strains section, where I'm going to post a list of the strains we've isolated over the years:
https://hotspringslab.ru/en/resources/strains
Each strain has its own card that tells you which medium to grow it on (recipes included) and at what temperature, along with some working notes.
For the published strains, I've also uploaded genome annotations.
You might reasonably ask why bother, since annotations can be downloaded from NCBI. Well, first of all, it's for my esteemed colleagues. Every now and then someone asks me something like "can you send me the HB27c sequence?", and of course I'm tempted to reply with something like "What, too much hassle to look it up in GenBank yourself?" But I know it's easier to just send it. Now I can answer with a clear conscience: "Have a look on our website." The site also lets anyone answer the question "How many plasmids does HB8 have?" (I've been asked this about five times lately), since each strain card includes a table listing its replicons.
On top of that, with the help of that same vibe coding, I've put together an all-in-one contraption that combines the outputs of several annotation tools:
- Bakta predicts ORFs, rRNAs, t(m)RNAs, and ncRNAs, and also searches protein sequences against a database;
- Baktfold makes structure predictions for unannotated proteins and searches them against structure databases;
- CCTyper predicts CRISPR-Cas systems;
- DefenseFinder and PADLOC predict other defense systems;
- geNomad predicts prophages;
- ISEScan predicts insertion sequences.
Also, CCTyper often gets the orientation of CRISPR arrays wrong. Since we know the repeats of the Thermus systems, I made a little thing that reorients the arrays and marks the boundaries between spacers and repeats.
And most importantly: Thermus genomes very often contain integrative elements that insert into tRNA genes. Existing tools either can't annotate these elements at all or do it poorly, so I bolted on a pipeline developed by Ilya Rubinstein that detects exactly these elements.
Admittedly, the annotations can look pretty cluttered, but please take a look and let me know what you think!