Exon Calling (based on Gencode v10 annotation):  Exons in Gencode V10 are quantified in this file.  Each exon gets a score representing the sum of the reads from both bioreps. The exons are then independently scored for each replicate according to their mapped read count as well as their unique barcode number.  The unique barcode at the 5' end of each read is used to remove the effect of redundancy in npIDR calculation. In stead of read count, unique barcode number is also used to compute Non-parametric-IDR (A Dobin, CSHL) values for each exon. In addition, the raw npIDR values are further smoothed using generalized polynomial curve fitting at degree 1. These results are reported in .GFF format.

.GFF Format:
Column 1: Chromosome ID
Column 2: customized label
Column 3: biotype of element (should be all "exon")
Column 4: exon start
Column 5: exon stop
Column 6: RPM 
Column 7: strand
Column 8: frame (should be all ".")
Column 9: contains additional information about the exon elements. "gene_type" means which what type of gene does the element belings to; "gene id" means which Gencode gene does this exon element belong to; "transcript id" means which Gencode transcript does this exon element belong to; "smIDR" means the smoothed IDR value for the exon.


Generated by: Wei Lin, CSHL

Contact: Carrie Davis (davisc@cshl.edu) and Wei Lin (wlin@cshl.edu)