Genome-wide identification and predictive modeling of tissue-specific alternative polyadenylation

Supplementary information for the ISMB paper by Dina Hafez, Ting Ni, Sayan Mukherjee, Jun Zhu, Uwe Ohler (2013)

Duke University, Institute for Genome Sciences and Policy

• Detailed explanation of the PA-seq protocol is available here
• All genes expressed in each tissue type (constitutive and alternative), in the following format: Gene,PAS,PAS_count,PAS_start,PAS_end
  • Brain
  • Liver
  • Kidney
• Code for our linear model: LM.R
• Input and output file for linear regression analysis (section 2.2: Characterization of tissue-specific regulated polyadenylation sites)
  Input file contain all alternative PAS for genes expressed in the three tissue libraries, along with their PAScount. Output file contains the residuals for each PAS used in each tissue
  • Linear model input
  • Linear model output
• Positive and negative datasets for constitutive sites used in the SVM classifier in seciton 2.3 (Modeling constitutive polyadenylation sites)
  • positive dataset
  • negative dataset
• Tissue-specific regulated PAS sites
  • Individual regulated PAS
    • Brain-specifc
    • Liver-specific
    • Kidney-specific
  • Overlap regulated PAS
    • Brain-Kidney-specifc
    • Brain-Liver-specific
    • Kidney-Liver-specific