ADpred documentation

Introduction

adpred is a tool to predict Transcription Activation Domains from protein sequences. It includes a command line script, methods and a class to simplify the commands.

For single predictions use to the adpred-web server

Adpred uses psipred to predict secondary structure of the sequences, which, together with the sequence conforms the input of the CNN model.

By default, adpred uses psipred server to calculate the secondary structure. However, the user can define (after loading the module):

local_psipred = <path/to/psipred_run>  # e,g, '~/some-path/run_psipred'

or, if run_psipred is loaded into the environment:

local_psipred_env = <command>  # depending on the version, In my case is 'run_psipred'

and adpred will use the local psipred installation.

The simplest use of the command-line tool is:

run-adpred --uniprot-id GCN4_YEAST

# to prompt a help message
run-adpred -h

# or just
run-adpred

The simple and quick use of the python module is:

# import the module
from adpred import ADpred as adp

# with own sequence:
myProtein = adp.protein("myProteinName", "ATREFERTATREFERTAADDWLNDCWATREFERTA")

# with uniprot identifier (example gcn4)
myProtein = adp.protein('GCN4_YEAST')

# if secondary structure is not known:
myProtein.predict()   # This will predict secondary structure

# If you wish to first get the secondary structure
myProtein.predict_second_struct()

# do saturated mutagenesis analysis to 30mer between residue positions 108 and 138.
myProtein.saturated_mutagenesis(start=108, end=138)

# By default the WT structure is used for all mutants, however,
# If you wish to recalculate the secondary structure for each mutant
myProtein.saturated_mutagenesis(start=108, end=138, 'second_struct_on_each_mutant')

Indices and tables

Credits

  • Ariel Erijman