I need to find homologs of certain proteins by BLAST search or similar. But I want filter (select) homologs from organisms that are classified as thermophilic or hyperthermophilic. I am wondering if there are suggestions for me how approach to this. Thank you.
It seems like the main issue is finding a list of species to filter your blast results on. I will therefore address this list of species issue.
There are some resources such as Microbe Directory that contain organismal characteristics similar to what you describe. Their main site appears nonfunctional but their github still works and appears to have the full database. This database includes an
extreme_environment field, which appears to be what you are going for.
I feel like PATRIC used to have something similar but I can't locate it now.
For more ideas you might look at the Microbe Directory paper to see if they cite other similar databases.
As the others have said, filtering by a list of taxonomy ids is the only one. However, there is one extra consideration if this is for wet lab applications: expression yields even when codon optimised may be affected by phylum choice.
Archaeal protein encoded by codon-optimised genes for E. coli generally perform less well than the equivalent bacterial thermophilic protein. I have not read a study on it, but I have heard its tale over and over and had issues myself. GC-content and the amino acid composition do correlate, but I have heard of good results in E. coli with Geobacillus (high-GC thermophilic firmicute) as with Bacillus (high GC mesophilic firmicute) after all. I don't know if its isoelectric points close to pH 7 or what. But mid GC-content bacterial thermophiles, such as Aquifex or some clades of Betaproteobacteria, may be a better call than Archaea.
So as you need to filter your blast search by a series of phyla, you may want to be selective if for expression in a different host...