The eqmr-fdb program creates a database for the phenotypic features. Several kinds of features are available:
- probe: to look at the probe expression levels;
- exon: to look at the exon expression levels;
- metaexon: to look at the metaexon expression levels;
- transcript: to look at the transcript expression levels;
- gene: to look at the entire gene expression levels.
When either an exon, a transcript or a gene has multiple probes inside, there will be a single final expression level per sample. To summarize the expression measurements from multiple probes, eqmr-fdb can use the following methods:
- mean: for each sample, compute the mean among the probes;
- median: for each sample, compute the median among the probes;
- medpol: apply a median polish algorithm on the sample x probe matrix, and use the row-effect estimates as the expression levels;
- pca: apply PCA (SVD) on the sample x probe matrix, and use the scores of the samples on the first PC as the expression levels.
For the exon or metaexon levels, it is also possible to compute specific exon levels by using the two following methods:
- firma: applied FIRMA method;
- e/g: scales the exon level by the gene level (sample per sample).
|-d||--data||The database file (.db)|
|-f||--feature||Either the feature type or the feature database (.fdb)|
|-m||--method||The summarization method for feature with multiple probes|
|-o||--output||The output file|
|-p||--min-probe||The minimum number of probes (default is 1)|
|-e||--exon-index||When -f exon, the method to create the exon index (default is none)|
|--gene-to-feature||Output the gene-feature relationship for a given feature database|
|--exon-covariate||For exon feature database, output details on exon features|
|--metaexon-covariate||For metaexon feature database, output details on exon features|
|--stats||Output Median/MAD of each feature within each population|
$ eqmr-fdb -d data.db -f gene -m median -o data_genexp
- As for eqmr-db, the created database will store the location of the corresponding directory. If you create the database by providing a relative path then you will have to run subsequent analyses from the folder you created the database. To avoid this issue, we deeply encourage you to provide to eqmr-fdb the absolute path in the option -o. In that case you could run the other programs of eQTNMiner in any places on your machine. Nevertheless, in that case DON’T MOVE the database folder to another location - for the same reason as mentioned previously. If you want to move the database folder, since the creation of the database is generally very fast, just rerun eqmr-fdb with the new location.
- Definition of a metaexon: a genomic region that will be converted into amino-acids in some, or all, of the transcripts (see Hoffman & Birney, Mol. Biol. Evol. 2007).