Is it possible a conda installation of psap?

[agg437]

Hello psap team!
I was wondering if it would be possible a conda installation of psap instead of using pip and how dependencies should be configured.
Best regards,
Ana

[agg437]

Hello again,
We tried to install psap environment with the following dependencies:
name: psap
channels:

• bioconda
• conda-forge
• defaults
• anaconda
• r
  dependencies:
• numpy
• biopython
• scikit-learn
• adjustText
• statsmodels
• pandas
• matplotlib
• seaborn
• psap
• r-essentials
• r-ggplot2
• r-ggpubr
  prefix: /home/gonzalezga/miniconda3/envs/psap

After executing the following command:
psap predict -f 'TIA1a prot EEE 383-385.txt' -o proof.csv
We got this error:

022-02-17 17:24:48.296 | INFO | psap.classifier:predict:151 - Loading model: /home/gonzalezga/miniconda3/envs/psap/lib/python3.10/site-packages/psap/data/model/UP000005640_9606_llps.json
2022-02-17 17:24:49.247 | ERROR | psap.classifier:predict:154 - classifier /home/gonzalezga/miniconda3/envs/psap/lib/python3.10/site-packages/psap/data/model/UP000005640_9606_llps.json not found. Does the file exist?
2022-02-17 17:24:49.281 | DEBUG | psap.classifier:export_matrix:38 - Starting to add biochemical features to TIA1a prot EEE 383-385.txt
2022-02-17 17:24:49.281 | DEBUG | psap.matrix:fasta2df:99 - Converting peptide fasta to data frame

Any ideas of how to solve it?
Thanks in advance

[tilschaef]

Hi @agg437 , we recently updated psap to v1.0.7 which should also be on bioconda once the version bump is merged. The new release contains the fix implemented by @hryk #28 .

Could you let me know if this resolves your issue?