The desired number of core orthologs could not be reached. Training with only 0 sequences

[guichenzhu]

Hello, I hope you don't mind me reaching out. I encountered an issue while using the software. My command was: h1s --seqFile infile.fa --seqName test2 --refspec PHYPA@3218@1 --blastpath ./output/blast_dir --weightpath ./output/weight_dir --searchpath ./output/genome_dir --cpu 1 --hmmpath ./. The species I initially selected for constructing the core ortholog set were:
Anthoceros angustus
Anthoceros punctatus
Chara braunii
Chlorella variabilis
Marchantia paleacea
Marchantia polymorpha
Physcomitrella patens
Rhodobryum roseum
Takakia lepidozioides

Initially, I selected nine species, but the following error occurred:
WARNING: The desired number of core orthologs could not be reached. Training with only 0 sequences.

Although the core ortholog set construction failed, the process still proceeded to identify homologous sequences. I then changed the species to:
Hypnum curvifolium
Physcomitrella patens
Syntrichia caninervis
Rhodobryum roseum

However, the construction still failed. My question is: Is this issue caused by the large evolutionary distances among the species selected for the core ortholog set? My study involves green algae, charophytes, liverworts, mosses, and hornworts. If that is the case, why does the problem persist even when I switched to using only moss species? Additionally, would it be acceptable to use the homologous sequences identified in the subsequent steps for phylogenetic tree construction, even if the core ortholog set construction fails?

[trvinh]

Dear @guichenzhu ,

Thanks for reaching out! :)

I believe that the core ortholog groups were not trained with zero sequences, but with a single sequence — the seed sequence. The warning message in this case is a bit misleading.

There are several possible reasons why the core compilation failed to find an ortholog in the core species. Technically speaking, this occurs when the sequence similarity is not high enough to meet the criteria for h1s to identify an ortholog. More specifically, either the HMM search in the protein set of the core species failed, or the HMM hits (ortholog candidates) did not pass the reciprocal BLAST search criteria.

Running h1s with a single CPU and enabling debug mode (using --debug and --silentOff option) will provide more information about the intermediate steps.

However, please note that we are no longer maintaining h1s and therefore cannot provide further assistance in investigating this issue. We recommend switching to its replacement, fDOG, which was published recently https://doi.org/10.1093/molbev/msaf120.

Best,
Vinh