TiSimilaritySNO

From Icbwiki

Jump to: navigation, search

S-nitrosylation target prediction with TiSimilarity

This page describes how tissue expression profile similarity searches can be used to predict targets of S-nitrosylations.

We use as training set the 68 proteins reported in Hao et al 2006. We search the human transcriptome with tiSimilarity as follows:

 tiSimilarity -b human-counts-data 
              --input H:\projects\gross\0DSNO\tisim\all-human-transcript-ids.txt 
              -o H:\projects\gross\0DSNO\tisim\sno-positive-human-screen.txt  
              --restrict  H:\projects\gross\0DSNO\tisim\sno-positive-human-transcripts-ids.txt  
              -k 100000 
              -a true  
              --scorer edu.mssm.crover.tools.tissue.similarity.scorers.ExpressionConfidenceScorer 
              -w true

Options indicated above mean:

--input file Each transcript ID listed in this file is used as a query transcript in turn. We use a list of all transcripts in the human genome.

-o where the result of the screen will be written.

--restrict the training set with known SNO proteins (one transcript ID per line for each protein in the training set). This option instructs tiSimilarity to sum scores and inverse of ranks only for those target transcripts that appear in the restrict list.

-k 100000 Accumulates scores and rank for up to top 100,000 transcripts. 100,000 is larger than the number of transcripts in the human genome, so all transcripts are considered.

-a true Accumulate scores and inverse ranks and prints only accumulated values in the output.

--scorer class Choose the confidence scorer.

-w true Determine tissue weights from transcripts in the restrict list. Weights are used when scoring to give more importance to those tissues where many restrict transcripts are expressed, and less to others.

Personal tools