Transcript Dell`Orphano: SNP discovery

Mining SNPs from EST Databases
Picoult-Newberg et al. (1999)
Strategy to Identify SNPs
• Used ESTs derived from 19 different cDNA libraries to assemble
300,000 distinct sequences and identified 850 mismatches.
• ESTs are short single pass cDNA sequences generated from randomly
selected library clones.
• EST contigs were generated using Phrap, a sequence alignment and
contig assembly program. (Picture of EST alignment)
• To distinguish between true SNPs and artifacts within the cDNA, a
series of filters were applied.
Filter Types Used
• Filter Test: 100 contigs were randomly selected and inspected with
Consed after filtration. Candidate SNPs were considered real if trace
data was of high quality and passed all four filters.
• Filter 1: Eliminates clusters of mismatches that often occur in regions
of low-quality trace data.
• Filter 2: Identifies sequence mismatches by either base substitution
type or insertion/deletion type.
• Filter 3 & 4: Addressed quality of each base call relative to its position
and frequency in a contig. First 100 bases discarded.
Consed view of a contig
containing a high quality
mismatch (A vs. T). The
mismatch has been confirmed
as a common SNP.
Filter Types Used
• Focused primarily on base substitutions (A/G or T/C) as these types
make up greater than 60% of all polymorphisms
Conclusions
• Successfully demonstrated a strategy for the rapid identification and
verification of SNP-based genetic markers using EST data sources.
• Possible that this approach could identify sequence variants that lead to
amino acid substitutions that may lead to functional differences.
• To show effectiveness of this strategy, the throughput of SNP
confirmation was increased by using Genetic Bit Analysis (GBA).
Overlapping Genomic Sequences: A Treasure Trove
of Single-Nucleotide Polymorphisms
Taillon-Miller et al. (1998)
SNPs in Overlapping Clones
• Developed strategy to identify SNPs in overlapping BAC clones.
(Picture)
• Sequenced three sets of overlapping clones, 153 polymorphisms (1 per
1.3 kb) were discovered to be unique (substitution, insertion or
deletion).
• 55 were discarded by computer analysis (filters).
• The oligonucleotide selection program (osp) was used to design
primers to amplify the 98 remaining SNPs.
• 30 SNPs were in regions with no suitable amplimers.
• In all, 44 STSs were developed to amplify the remaining 68 SNPs,
with 16 STSs containing 2 SNPs and 8 STSs contained 3 or more
SNPs. All 68 SNPs were present in at least one of three populations
studied.
SNPs in Overlapping Clones
30 individuals
Conclusions
• Informative SNPs can be found by using overlapping regions of clones
because BACs are typically from different individuals.
• They described a marker every 4.8 kb.
• Many of the SNPs have different population frequencies
• This approach has many advantages because 1) high quality sequence
data because every base in overlap is sequenced at least twice; 2) SNP
data is generated by analyzing existing data; 3) SNPs are derived from
long range sequence data and these markers are precisely mapped.
• SNP markers can be found in overlapping genomic sequence, it is
highly efficient and cost effective. Only two steps involved (develop
STSs around known SNPs and characterize them in populations).
An SNP map of the human genome generated by
reduced representation shotgun sequencing
Taillon-Miller et al. (1998)
Reduced Representation Shotgun Sequencing
• RRS re-samples specific subsets of the genome and compares the
resulting sequences using a highly accurate SNP detection algorithm.
• Prepare subsets of the genome (reduced representation), by performing
restriction digest to generate restriction fragment lengths between 500
to 600 bp. (Picture)
• Computer analysis of 517 megabases yielded 3847 BglII fragments.
• Thus SNPs can be discovered by mixing DNA from many individuals,
prepare library of sized restriction fragments, and randomly
sequencing clones.
• Defined a neighborhood quality standard (NQS) to cut down on base
calling errors. Good sequence results in higher accuracy.
Reduced Representation Shotgun Sequencing
• Used polymorphism free BAC DNA to evaluate NQS. Any ‘SNPs’
would represent base calling errors.
• Base calls within ‘good neighbors’ were more accurate than predicted
by Phred.
Conclusions
• Demonstrated a new method for re-sampling loci without PCR, and for
detecting SNPs with high accuracy in the resulting alignments.
• RRS can be configured for any organism.
• SNPs discovered by RRS offer the potential for reduced representation
genotyping without locus specific amplification.