README.md

In dnusinow/snipeR: Software for Network Inference of Proteomics Experiments

Installation

Installing the R Packages

snipeR is provided as an R package. This may be installed using the UNIX shell command

R CMD INSTALL snipeR_0.7.6.tar.gz

You may need to use the -l option to specify where R installs the package to. See the manual R Installation and Administration that comes with the R distribution. You will also need to have installed several external R packages including:

• Matrix (standard with R)
• foreach
• optparse
• stringr
• doMC
• RSQLite

We recommend using R's install.packages() function for this. If you aren't familiar with this function, simply copy and paste the following line of code in to your R session and after asking you which mirror to use it'll install the packages for you:

install.packages(c("Matrix", "foreach", "optparse", "stringr", "doMC", "RSQLite"))

Installing the Necessary Databases

Once you have installed the package you need to download the necessary database files. The easiest way to do this is to use the download.missing.maps() function provided with the package. Provide this with a single optional path to where you want the downloaded files to go and it will do the rest. Beware, some of these files (most notably STRING) are quite large and can take a very long time to download. If you would prefer to download these files by hand, you will need:

• protein.links.v8.x.txt.gz and protein.aliases.v8.x.txt.gz from the Downloads section of the STRING homepage. For the paper we used version 8.2, which can be found at http://string82.embl.de/

• idmapping_selected.tab.gz from UniProt.

If you do download these files by hand then you'll need to run the string.aliases.to.sqlite() and string.raw.to.sqlite() functions to process these files in to SQLite databases. Each of these functions requires only a single parameter: the location of the downloaded files to be processed (either the path to the file itself or the directory it is housed in.) You will need permissions to write to this directory, as that is where the SQLite database is created. Please see the documentation for these functions, included in the snipeR package as standard R help files, for more detail.

Usage

This is more fully documented in the R help for the snipe() function. snipe() requires both a treatment and control data frame for comparison, which can be passed directly to the function as parameters. Alternatively, you may supply snipe() with paths to tab-delimited text files for the treatment and control samples containing the spectral counts and protein ID's. Ensembl Protein IDs (STRING's native encoding) and IPI ID's are currently supported. You will also need to supply a species taxonomy ID (i.e. 9606 for humans, 10090 for mouse) and a path to the location of the downloaded files listed above. Other useful parameters include the path to the location of various downloaded and processed database files (see Installation above).

To adjust the sensitivity of the calculated p-values, you may supply a number for permutations to be performed for each protein. Additionally, you may speed performance on multicore UNIX machines by supplying a number of jobs to split the load into. The number of permutations is per-job, so if you want a million permutations split across 10 CPUs, use 10 jobs and 100,000 permutations per job. This will not currently work across a cluster, only on a local multicore machine.

The snipe() function returns a data frame that contains the output of interest. This data frame can easily be written to disk using a standard R command such as write.table(). We provide a shell script named snipe in the snipeR directory that can be used as an example for how to use the package. This has not been extensively tested, and so should be primarily considered as a usage example.

Note also that we have not completed porting the program to Windows, nor have we tested it on a MacOSX system. Although snipeR should theoretically work on both of these systems, we can only say that it works on Linux with any reasonable confidence right now.

Reproducing The SNIPE Paper

For the paper, the spectral counts for each replicate in each class were summed to generate specific Tooth Germ and Oral Tissue files that were then passed in to snipeR. To create these files easily we recommend R's merge() function to match up the IPI ID's followed by some simple summing and cleaning of the data. However, to speed this process we have included these data in the snipeR distribution as data frames that may be loaded in to R using the data() function passing it name Tooth_Proteins. See the examples included in the help documentation for the snipe() function for code to reproduce this work.

dnusinow/snipeR documentation built on May 15, 2019, 9:40 a.m.