output tree won't write

[qiyunzhu]

Hello @scapella ,

PS: Sorry, I think I didn't understand the pipeline correctly. The final output tree was renamed from kinase.alg.clean_phyml_tree to kinase.tree.phyml.ml.<model>.nw. The best tree is the first line of kinase.tree.phyml.rank.ml. Is that correct?

==========================

I am playing with Phylomizer. So far I noticed one major issue, that the final output tree was generated but not saved in the output directory. Can you please take a look? Thank you!

I started with a collection of protein sequences (kinase.fa), and ran --steps alignments trees. It seems that the task was completed successfully:

-bash-4.1$ tail -n 20 output/kinase.log 
. Better tree found so far has lnL:    -12993.23 (4 more trees to examine)
. Better tree found so far has lnL:    -12986.98 (3 more trees to examine)
. Better tree found so far has lnL:    -12984.91 (1 more trees to examine)
. Improving the best trees using NNIs...
. Better tree found -> lnL:    -12981.42 (5 more trees to examine)
. Checking for NNIs, optimizing five branches...
. Log likelihood of the current tree: -12981.373259.
. Calculating fast branch supports (using 'aLRT').
. Printing the most likely tree in file 'kinase.alg.clean_phyml_tree'.
. Time used 0h9m14s
oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo
### Time    0d:00h:09m:14s
###
###
### STEP    Phylogenetic Tree Reconstruction    END 11:12:57 02/09/17
### TOTAL Time  Phylogenetic Tree Reconstruction    0d:00h:23m:55s
###

### TOTAL Time  [ 'alignments', 'trees' ]   0d:00h:25m:22s
###

However, file kinase.alg.clean_phyml_tree is not present in the output directory. My configuration file is:

verbose             parameter    1
residue_datatype    parameter    protein
e_value             parameter      1e-5
coverage            parameter      0.5
hits                parameter      150
force_seed_sequence parameter      True
homology            mode           blast+
blast+              mode           psiblast
psiblast            binary
psiblast_params     parameter      -seg yes -num_alignments 5000 -num_threads 2
psiblast_params     parameter      -outfmt 6 -use_sw_tback -dbsize 1000000
psiblast_params     parameter      -comp_based_stats 0
alignment           mode           kalign muscle mafft
consensus           mode           m_coffee
trimming            mode           trimal
both_direction      parameter      True
min_seqs            parameter      3
in_letter           parameter        U:B
in_letter           parameter        O:Z
muscle              binary
muscle_params       parameter
mafft               binary
mafft_params        parameter       --auto
kalign              binary
kalign_params       parameter -f fasta
m_coffee            binary           t_coffee
m_coffee_params     parameter        -n_core 1 -output fasta -quiet
trimal              binary
trimal_params       parameter         -phylip -gt 0.1
trimal_compare      parameter         -ct 0.1667
readal              binary
tree                mode           phyml
evol_models         parameter        JTT WAG MtREV VT LG Blosum62 Dayhoff
numb_models         parameter        2
tree_approach       mode           nj ml
nj                  parameter        -b  0 -o l
ml                  parameter        -b -2 -o tlr
phyml               binary
phyml_params        parameter      -d aa -f e -v e -a e -c 4 --no_memory_check

The same thing happened when I tried RAxML instead of PhyML. It finished successfully but the tree files are just not there.

[scapella]

Yes - you understand the pipeline correctly. To ensure uniformity across programs and versions, tree files are renamed to .tree....nw & .tree....st.

In your case, files will be renamed from kinase.alg.clean_phyml_tree to kinase.tree.phyml.ml..nw for the ML approach.

[qiyunzhu]

@scapella Thanks for the clarification!