Hello all,
I am using this command phylophlan -i /mnt/bigdata/nikhil/Bradyrhizobium_genomes -d phylophlan --databases_folder /mnt/bigdata/nikhil/Bradyrhizobium_genomes/database --diversity medium --configs_folder /mnt/bigdata/nikhil/Bradyrhizobium_genomes/phylophlan_configs -f supermatrix_aa.cfg --nproc 80 --verbose to do phylogenetic analysis on my 156 bacterial genomes. However, it’s been around more than 138 hrs and analysis is still going on. I checked the steps, there is no error in the analysis. Kindly provide guidance.
Hello @Rishav_Sahil, thanks for using PhyloPhlAn and reporting this. It’s strange to me that it will take that much time for 156 bacterial genomes. Since you’re running with the --verbose, will you be able to share the current output form PhyloPhlA? Or, even better, if you’re redirecting that to a log file and can share it, I can try to have a look at the steps and their runtime.
Thanks,
Francesco
Hello @f.asnicar , firstly thank you for your response. For alternative, i ran it on the proteomes output by prodigal tool and for that it works fast and gave me the result. However, I didn’t stop the analysis on genomes as i wanted to see how much will it takes to complete it. After 483 hrs of running, it still needs to generate 3 more .bkp files out of 156 and i think it will now complete within one to two days.
grep "generated in" phylophlan_run.log | grep -oP '"\K[^"]+(?=.b6o.bkp")' | sort -u | wc -l
153
The most probable reason is that maybe some draft genomes are highly fragmented.
Hello @Rishav_Sahil, thank you for confirming this. Indeed, the translated mapping is a very slow process (unfortunately), and that depends on the tool (which is external to PhyloPhlAn) and the characteristics of the input genomes. Definitely, running prokka/prodigal/bakta and using the proteomes as input would have been my suggestion, as direct amino acid-to-amino acid mapping is much faster.
Thanks,
Francesco