Code and data for the SIEVE-Ub prediction algorithm for prediction of E3 ubiquitin ligase mimics
Accompanies the preprint: https://peerj.com/preprints/27292/
Published paper: https://peerj.com/articles/7055/
The model used for the prediction script was not effective at classifying the training examples. This doesn't reflect the performance of the method, which is carefully validated in the paper and is based on separate analysis that doesn't rely on the model used in the final prediction script.
I've updated the script to use the corrected model which provides 94% correct classification of positive examples and 99% correct classification of negative examples.
Files:
KmerFeatures.py : Python code that transforms a protein fasta file into kmer features suitable for use
with. the SIEVEUb.R code.
SIEVEUb.R : R code to train and validate models based on kmer features. Includes code for family-wise
cross-validation and family-wise scoring of overrepresented kmers.
SIEVEUbApply.R : an R script that allows reading in a set of features output from KmerFeatures and applying
the predictive model from the paper to output predictions for each protein.
SIEVEUbCode.Rmd : RMarkdown file for doing the analysis described in the paper. This RMarkdown has code for
regenerating most of the figures, tables, and supplemental figures and tables from the paper.
make_prediction.csh : shell script that will provide predictions of ubiquitin ligase given an input fasta file.
data/1032069.3.PATRIC.faa : example PATRIC genome fasta file from a single genome
data/FamiliesConservative.txt : list of example proteins used for positive and negative examples, their annotations,
and identified sequence families
data/PATRIC.genomes : list of genomes from PATRIC used in the paper
data/PATRIC_secretion_systems.txt : Type III, IV, and VI secretion system components identified in each PATRIC genome used
data/SIEVEUb_best_model.Rdata : Rdata file containing the best Ub ligase model
data/SIEVEUb_minimal_model.Rdata : Rdata file containing the minimal Ub. ligase model with 10 features
data/best_model_1280features.txt : list of features included in the best Ub ligase predictive model
data/legionella_examples.fasta : FASTA protein sequence for newly discovered UUb ligase RavN
data/legionella_examples_k14red0_minimal10.txt : feature file for RavN using the minimal model identified
data/minimal_model_10features.txt : kmers contained in the 10 kmer minimal model
data/patric_ubligase_predictions_best_model.txt.gz : predictions for the PATRIC subset used in the paper with the best
Ub ligase predictive model
data/ubligase_examples_ids.fasta : FASTA file containing protein sequences of positive and negative examples used in this
study. Please see FamiliesConservative.txt for information about these sequences.
data/ubligase_examples_ids_minimal_model_10features.html : HTML file that lists the examples used in this study and where
the top 10 most predictive kmers map to the sequences.
data/ubligase_examples_ids_minimal_model_10features.tab : List of most predictive 10 kmers locations in each of the examples
used in the study.
data/uniprot_search.readme : short description of Uniprot search and filtering process used to identify positive
examples for E3 ubiquitin ligase
data/red0/ : features for reduced alphabet 0 over a range of kmer lengths from 3-20
data/rnd0/ : features for randomized binary alphabets
This material was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the United States Department of Energy, nor Battelle, nor any of their employees, nor any jurisdiction or organization that has cooperated in the development of these materials, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness or any information, apparatus, product, software, or process disclosed, or represents that its use would not infringe privately owned rights.
Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof, or Battelle Memorial Institute. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.
PACIFIC NORTHWEST NATIONAL LABORATORY
operated by
BATTELLE
for the
UNITED STATES DEPARTMENT OF ENERGY
under Contract DE-AC05-76RL01830