Computational Analysis of Transcript Identification Using GenBank Slides by Terry Clark
Differentiation of hematopoietic cells
Genome-wide gene expression
SAGE (Serial Analysis of Gene Expression)
Jes Stollberg et al. Genome Res. 2000; 10: Figure 1 Schematic illustration of the SAGE process
SAGE & GLGI Overview
What is the chance of duplicate tags? We can assume we are drawing randomly from the set of all 4-letters sequences of the given tag length This is the same problem as having unique overlaps in the contig matching problem for shotgun sequencing
Random Model
Random model does not reflect biological process Genes evolve by duplication as well as point mutation Many motifs are repeated Function widgets at work? Result is a strong bias in observed biological sequences, not a uniform distribution as the simple model hopes. Here are some numbers ….
SAGE tags match to many genes (Tags from Hashimoto S, et al. Blood 94:837, 1999)
Tag Frequency Groups for 10-base Tag Set Containing 878,938 Tags for UniGene Human
Unique Tags among 878,938 EST Derived Tags
Unique Tags among 32,851 Gene Derived Tags
Converting tag into longer 3’ sequence
Generation of Longer 3'cDNA for Gene Identification (GLGI)
UniGene Human 3’ Part Length Distribution
Myeloid Tag Matches with UniGene Human SAGE Tag Reference Database
SAGE Tag Processing with GIST
k-mer tree
GIST Performance with Improved IO
Conspirators Sanggyu Lee Janet D. Rowley San Ming Wang Terry Clark Andrew Huntwork Josef Jurek L. Ridgway Scott