Skip to content

Latest commit

 

History

History
54 lines (40 loc) · 2.66 KB

README.md

File metadata and controls

54 lines (40 loc) · 2.66 KB
Raw

ProteinPatternSearch

This repository includes Java classes that do the pre-processing steps of filtering transmembrane protein fragments from the SwissProt protein database. Project details can be read in Chapter 5 of my dissertation

The tasks this software package is able to perform are:

  • COMPUTE_STATS: Expects a Swiss-Prot .dat file and counts ’H’, ’D’ and ’HD’ frequencies, and prints an expected ’HD’ frequency, assuming an uniform and independent random distribution of amino acids.
  • FILTER_DAT_ROWS: Expects a .dat file and codes of Swiss-Prot lines to throw out; writes a shorter .dat file.
  • FASTA_AND_DAT_FILTER: Filters both a .dat and a .fasta file, keeping only proteins that match a pre-defined pattern of keywords and other features, as well as an optional minimal sequence length.
  • TM_STATS: Prints statistics of transmembrane proteins of a .dat file (number of total proteins, number of TM proteins, and number of TM proteins with extra- or intracellular annotation).
  • TM_EXT_FRAGMENTS: Expects a .dat and a .fasta file, writes a .fasta file with extracellular protein fragments and the neighboring transmembrane sections. The minimum and maximum fragment lengths can be specified via the minLenOfExtracellularPart and maxLenOfExtracellularPart properties.
  • SPLIT_FASTA_TO_CLUSTERS_KCLUST: Reads kClust output files and a .fasta input; writes many .fasta files corresponding to the clusters. Discards too small clusters, if the minClusterSize parameter is set.
  • STATS_ARFF_FOR_WEKA: Generates an .arff file (the input format of Weka) from a .fasta file of proteins or protein fragments; the attributes (columns of the .arff) will be the different amino acid counts, or amino acid group frequen- cies for each protein(fragment).
  • WEKA_CLUSTERING: Runs k-means clustering, built into Weka. Besides the input and output paths, mandatory parameters are: numberOfClusters and randomSeed.
  • SPLIT_FASTA_TO_CLUSTERS_WEKA: This task is very similar to the task "SPLIT_FASTA_TO_CLUSTERS_KCLUST", but it works with Weka output files instead of kClust results.

How to use

A) Compile

To compile the java project, first get Gradle (version 1.6 or above), and the run:

gradle build

B) Run the whole pipeline

The easiest way to get a working example is to follow the scripts and configs as described in my dissertation. There you will see where are the lines in the configs that should be changed to execute the programs on your files, with your parameters.