Dear All,
I have a doubt in the calculation of False postitive rate while checking for FPKM threshold in a RNAseq experiment.
Following the method previously published (http://www.ploscompbiol.org/article/...l.pcbi.1000598). I am not getting desired results.
I followed the method as mentioned the publication
Reads were mapped to Ensembl genes (blue) and intergenic background regions (red). The intergenic regions were matched to have the same length distribution and no ESTs mapping within regions. We binned the expressions of all genes and background regions across human tissues. We zoomed in on the effects at expression spanning between 0.01 to 10 RPKM for clarity. Bins were converted to cumulative amounts of genes expressed above the expression levels for genes and controls cum_background; A false discovery rate fdr was calculated at each expression level, i, as:
fdri=cum_backgroundi/cum_genesi * (1-cum_genesi)/(1-cum_backgroundi).
I am confused over the fact that cum_background and cum_genes stand for the cumulative count of genes or they are the fraction of cumulative gene counts falling under each bin.
I have a doubt in the calculation of False postitive rate while checking for FPKM threshold in a RNAseq experiment.
Following the method previously published (http://www.ploscompbiol.org/article/...l.pcbi.1000598). I am not getting desired results.
I followed the method as mentioned the publication
Reads were mapped to Ensembl genes (blue) and intergenic background regions (red). The intergenic regions were matched to have the same length distribution and no ESTs mapping within regions. We binned the expressions of all genes and background regions across human tissues. We zoomed in on the effects at expression spanning between 0.01 to 10 RPKM for clarity. Bins were converted to cumulative amounts of genes expressed above the expression levels for genes and controls cum_background; A false discovery rate fdr was calculated at each expression level, i, as:
fdri=cum_backgroundi/cum_genesi * (1-cum_genesi)/(1-cum_backgroundi).
I am confused over the fact that cum_background and cum_genes stand for the cumulative count of genes or they are the fraction of cumulative gene counts falling under each bin.