search

cruizperez / MicrobeAnnotator

Pipeline for metabolic annotation of microbial genomes
Artistic License 2.0
137 stars 27 forks source link

stopped at step 14 #37

Open AhmedElsherbini opened 2 years ago

AhmedElsherbini commented 2 years ago

Hi guys,

Thanks for the good work, Using a cluster computer, I was downloading your databases the full version and it stopped at step 14

2021-10-30 20:29:18,929 [INFO]: Step 14 2021-10-30 20:29:18,929 smp01 main: Function: database_duilder - Line: 305 [ERROR]: Search method not recognized. Must be blast, diamond, or sword. Any clue what should I do ? Update: the mistake is related to the choice of which search method (choose only one method not the three ), so this is an example for the right command.. $ microbeannotator_db_builder -d MicrobeAnnotator_DB -m blast -t 12 --no_aspera

rotheconrad commented 2 years ago

Hello.

Did you get this working?

AhmedElsherbini commented 2 years ago

HI Roth,

Thanks for your reply,

Yes, it is working, I also try it on in-silico amplicons and it working. My observation is the tool makes a nice annotation tsv. To make sure that the annotation is "reliable" especially with kofam. do you just dependent on "E value" to assess the annotation ? for example, I see a discrepancy between Prokka annotations and your prediction, How would you make sense? My second observation, the metabolic graphs are really great but too many colors in the legend makes it a bit hard to be understood? could you like ordered descendingly ?

rotheconrad commented 2 years ago

The database Prokka uses is different from the databases used by MicrobeAnnotator. The common practice for annotation is to return the best match to the database. If the database is different the best matches can be different.

That is a good suggestion for the plot, but we are busy working on other things. MicrobeAnnotator provides the data files that the heatmap is generated from. You could use this to generate your own heatmaps in Python or R.