Genomics Studies - Biomedical Informatics in Translational Research - page 62

However, the power of SNPs lies less in their ability to predict changed protein
sequences and more in the fact that they are (1) numerous, (2) ubiquitous, (3) stably
inherited, and (4) very well characterized. At the time this topic was written, the Sin-
gle Nucleotide Polymorphism Database (dbSNP; http://www.ncbi.nlm.nih.gov/
projects/SNP) release 127 contained 11,825,732 validated SNPs and, therefore, pro-
vides a fairly evenly distributed set of markers across the entire human genome. For
this reason, one aim of biomedical informatics is to find a link between a SNP and a
particular disease symptom or outcome, thus identifying genetic loci that are linked
to a specific clinical trait even though the mutation giving rise to the SNP per se may
not cause changes in gene expression (e.g., it may occur in a noncoding part of the
gene or may even lie in an adjacent, unaffected gene). Hence, the SNP is a stable
genetic marker that may not necessarily be causative of the disease, but provides an
indication of an adjacent locus that is. This ability to link SNPs to a genetic locus led
to the establishment of the HapMap project in which SNP distribution was assessed
for five different ethnic groups [17] and has since been used to find relationships
between SNPs and various diseases [18, 19].
Another important event in gene expression is the processing of mRNA. As
described in Chapter 4, after transcription the precursor mRNA is processed and a
large percentage of it is discarded. This processing follows specific rules (most of
which are well characterized; e.g., splicing predominantly takes place between
AGGT sequences that flank the 5' and 3' ends of the intron) and this results in the
removal of variable lengths of precursor mRNA, which are not important for the
function of the coded protein. However, it is possible for one gene to give rise to
multiple proteins through a process known as alternative splicing in which different
exons are mixed and matched to provide similar proteins with slightly different
functionalities. There are many reports in the literature of such events [20]. Because
this is controlled by the DNA/mRNA sequence, alterations in the DNA sequence can
lead to altered splicing and production of proteins with altered functionality. For
instance, if the AGGT splicing signal is mutated, the cell may well find an alternative
(cryptic) splice site and use this instead, giving rise to a protein with altered function-
ality and so giving rise to disease [21-24].
Because these SNPs are well characterized, it is possible to build a chip that
screens for these SNPs (currently a 500K SNP chip is commercially available). The
following is true for SNP chips and is generally true for other chip-based technolo-
gies described later.

5.2.2.3 Affymetrix Genotyping SNP Assay Workflow
Affymetrix named their genotyping SNP array assay process whole genome sam-
pling analysis (WGSA). It includes a few key steps: restriction enzyme digestion,
adapter ligation, PCR amplification, fragmentation, and hybridization (Figure 5.1).
The Affymetrix protocols start with 1 g of DNA that is divided into four tubes

(250 ng of DNA per tube) for the restriction enzyme digestion, which cuts genomic
DNA into different sized fragments that allow the specifically designed adapter to
bind. After heat inactivating the restriction enzymes, adapter ligation is performed
using T4 DNA ligase. Specific primers for the PCR amplification are embedded in
these adapters. After another heat step to inactivate the DNA ligase, a specific PCR

μ