prodigal-gv 2.9.0 vs prodigal 2.6.3

[lonestarling]

apcamargo
Hi,
First of all, thank you for you excellent tools, i did some test on my virus metagenomic data by prodigal-gv and prodigal 2.6.3 in " p meta", and i finally get 2440739 proteins by prodiga-gv and 2496415 by prodigal 2.6.3, a little bit different in them. acctually, i don't know how to choose them.

And in the study " Metagenomic compendium of 189680 DNA virues from the human gut microbiome" by Stephen Nayfach, they have choosen the code (4) TGA and code(15) TAG and code (90) TAA besides the standard code (11), and then chooe the results accoding the potential score for every predicted gene. but't i haven't understand how to change the code in meta mode. and i don't how how much it changes the outcome in this way compare to the code 11 only.

So can you give me some advice.
Sincerely yours
L

[apcamargo]

Hi @lonestarling. In that paper Stephen did a manual selection of the genetic code by specifying the translation table and running Prodigal three times (one for each code), in the isolate mode, not metagenome. The goal with prodigal-gv is to automate this process by adding gene models with alternative codes to the metagenome mode.

The raw number of proteins is probably not a good comparison here, as standard Prodigal could be calling multiple small fragments for phages with alternate codes. Try to compare the code density of the gene calls generated by both software.

[lonestarling]

Hi
apcamargo
thanks for your patiently reply， i get what you mean. You must be an expert in this field.
Actually, i am freshman from china in field of DNA and RNA virus. recently, i am doing some study about gut phage of mammals.
You must know that taxonomy of viruses is a tricky issue. as the same time, according to the Stephen's paper, i can only do the step about diamond blastp for protein from NCBI Genbank, but i don't konw how to to further process these results into taxonomy results.
I got a lot of inspiration from you method about MMseq taxonomy. and have finish that with my data. it's amazing.
Look forward to communicate with you. I will really appreciate if i can get your email.
Best wishes
L