Transcript Identification of alternative spliced mRNA variants related to(1).

Oncogene. 2004 Apr
Identification of alternative
spliced mRNA variants related
to cancers by genome-wide
ESTs alignment
KIM DAE SOO
MPL
Alternative splicing Form
BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
Abstract
Alternative splicing of mRNAs by analysing the
exon linkage relationship by alignment of ESTs to
the genome sequence
 Little effort has been made to investigate the
relationship between cancers and alternative
splicing
 Alternative splicing assembler(ASA)
 Of 4322 genes screened,3498(81%) were observed
with at least one alternative splicing variants.
 Using Fisher’s test, alternative splicing variants

BIOINFORMATICS
MPL
Introduction





About 45% of the human EST sequence are derived from
cancer cells.
Alternative splicing is an important mechanism in higher
eukaryotes for producing proteomics complexity
Approximately 30-60% genes are alternative splicing as
estimated by genomically aligned ESTs
Human could conceivably produce hundreds of thousands
of different proteins by the estimated 35,000 genes
Alternative splicing of pre-mRNA is a versatile mechanism
for regulating gene function at the post-transcription level
BIOINFORMATICS
MPL
 Alternative
splicing defect as indicated by a
survey of mutations in splicing junctions.
 A total of 26812 alternative spliced variants
from 4322 genes were included in the
database
 2149 variants from 1827 genes were
predicted as cancer associated.
BIOINFORMATICS
MPL
Results
ASA was created to identify alternative splicing of human gene
transcripts
52 genes from alternative splicing database ,all
having previously reported alternative splicing
isoforms.
 Only four EST hits were available on the average
for these 6 genes ,202 EST hits could be found for
the other 46 genes
 ASA gives a false negative result of 12 %
(about 88% of alternatively splicing genes)
 SpliceNest,PALS,HASDB which have false
negative rates of 15,35 and 38%

BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
EST sequences might be contaminated by genomic
sequence, vector sequences and chimaeric cDNA
clone
 Randomly selected 82 splicing variants from 67
genes compared these variants with the NR db
 26%(21 out of 82) variants found at least one
reported mRNA sequences
 selected 13 variants ;RT-PCR
(11 of 13 variants were confirmed by RT-PCR)

BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
Analysis of the alternative spliced variants
and their tissue distribution in BASD
A total of 4322 reference sequences were screened
and 3490(81%) reference sequences were predicted
as alternative spliced ,producing 26,812 splicing
variants.
 On average six splicing variants were observed for
each reference sequences.
 87% variants contained less then 16 EST ,and 58%
were represented by the EST singletons.
 If the singletons were excluded from the database,
then only 66% genes were alternative spliced

BIOINFORMATICS
MPL
 A total
of 6593 libraries were left and
classified into 293 tissue types (NCBI EST
data)
 A sample of 127 random genes was used to
analysis the tissue distribution of alternative
splicing
 Large fraction of splicing variants are tissue
specific
BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
Identification of cancer-associated splicing
variants
 about
35% splicing variants were detected
exclusively in cancer tissues
 29% were only detected in normal tissues
 This implies that new splicing variants might
be generated during carcinogenesis
 This result supports the observation that
during carcinogenesis not only is the
expression profile affected but the splicing
Patten
BIOINFORMATICS
MPL
 This
383 genes of variants will have a
relatively higher confidence of cancer
relationship.
 About 85% cancer related variants had more
than 16 ESTs
 About 43% of cancer specific variants were
related to more than two kinds of cancer
types
BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
BIOINFORMATICS
MPL
Discussion
 We
constructed BASD to search for splicing
variants of gene transcripts especially those
associated with cancer
 These single EST supported variants in
mining genomic information
 Our work presents a genomic view of the
relationship between cancer and alternative
splicing.
BIOINFORMATICS
MPL