Every human cell, with two sets of 23 chromosomes, contains six-billion basepairs of DNA (or three-billion per haploid genome). Of these three-billion genomic basepairs, each individual shares 99.7% with the rest of the humanity. It is the three-tenths of a percent that determines the differences between all of us. This tiny percent, nevertheless, comprises of about a million positions that not only make us unique individuals, but also determine how we respond to environment, succumb to certain diseases, or respond (or not) to certain drugs. These single nucleotide changes, scattered all over the genome, are called single-nucleotide polymorphism (SNP, pronounced snip) - for example, I may have Adenine at position X, you may have C and my friend may have G at the same position. Since the complete sequencing of human genome in 2003, the post-genomic goal has been, to answer how this 0.3% of genome determines phenotype. Pharmacogenomics/Pharmacogenetics (PGx) is the study of how genetic makeup correlates to responses to various drugs.
The promise of PGx is to match patients and drugs - the concept of "one size fits all" is slowly being thrown out of the window in the slowly creeping era of personalized medicine. Two main drivers of PGx are: getting the most effective drug to a patient and not using drugs that are likely to cause adverse events in a patient.
Russ Biagio Altman, M.D., Ph.D., Professor of Bioengineering, Genetics, and Medicine (and Computer Science, by courtesy), is the director of the Center for Biomedical Computation at Stanford University and is director of the biomedical informatics training program. He is also the principal investigator of an NIH project, PharmGKB (PHARM-gee-kay-bee), which is an online database of genetic and phenotype information from people who have participated in research studies at various medical centers participating in the NIH Pharmacogenetics Research Network (PGRN). In today's grand rounds, I'm attending his presentation at Google TechTalks series, he gave on February 22nd, 2006. Russ introduces the concept of PGx, the public domain database, PharmGKB, and introduces looming issues regarding patient privacy. One goal of PharmGKB is to correlate genomic information to molecular & cellular phenotype and to clinical phenotype. The genomic information is linked to literature, drug, PubChem and other databases; hand annotation of literature data, etc., makes it a very reliable and respected source. It also has pathway maps for drugs.
- Russ gave two examples demonstrating the power of PGx. The first was the differential metabolism of codeine by CYP2D6 variants. Codeine, an opioid, is metabolized to active molecule, morphine, in the liver by enzyme CYP2D6. Seven percent of Caucasians do not respond to codeine; they have a variant version of CYP2D6 that does not metabolize codeine. The second example was the story of clinical development of Bidil (BAI-dil). Bidil is a combination of two drugs, and was deemed a statistical failure in a phase III trial for heart failure. A sub-group analysis revealed positive response in people of African-descent; after another trial, FDA approved the drug for use in this particular subset of population. However, the color of skin is not a PGx or biomarker, but eventually we will progress to hard SNPs or genetic markers.
- He described the challenges to bringing PGx in routine use: limited data in public domain, fragmentation of drug markets; the ethical and privacy issues. In U.S., it is unclear how to deliver this information to the practitioner - interestingly, he said, countries like U.K., Canada, Estonia or Iceland, are all way ahead of us; they have centralized patient and drug information system that a doctor can access in the examination room from his computer terminal.
- In the Q&A, Russ expanded on the potential of risk to patient privacy. He said, among the 0.3% of unique DNA, just 60-100 SNPs may make a person unique. This is the "fingerprint" of a person. Combining this key SNP information with phenotype data could easily (potentially) unmask the anonymous nature of the database. This is a real concern. He went on to describe a real case of a UK teen who tracked down his sperm donor father at age 15.
"The boy in the New Scientist report tracked down his father from his Y chromosome, which is passed down from father to son. His genetic father had never supplied his DNA to the website that they boy sent the swab to - FamilyTreeDNA.com - but two other men who were on the database had Y chromosomes that bore a close match to the boy's. Both men had the same surname, although with different spellings. The two did not know each other, but the similarity between their Y chromosomes suggested there was a 50% chance that all three had the same father, grandfather or great-grandfather. Though the 15-year-old boy's donor had been anonymous, his mother had been told the man's date and place of birth and his college degree. The boy fed this information into another online service, Omnitrace.com, to find out the names of everyone that had been born in the same place on the same day. Only one man had the same name that he was looking for, and within 10 days he had made contact with this man." (BBC News: Boy tracks his sperm donor father)
- The story of the UK teen raises concern about privacy; anyone willing to link "publicly-available" information in various databases, can fingerprint an individual. No wonder, U.S. still lags in the adoption of PGx for routine use, although progress in areas such as, biomarkers, tumor profiling and cancer-drug matching has moved forward.
PharmGKB database today
PharmGKB website states the mission, "To collect, encode, and disseminate knowledge about the impact of human genetic variations on drug response. We curate primary genotype and phenotype data, annotate gene variants and gene-drug-disease relationships via literature review, and summarize important PGx genes and drug pathways"
What can we do with this wonderful database?
As an example, I typed in bevacizumab in the search box on top of the page.
The database returns two sets of results: database and web searches. The database provides information on the drug, its trade & generic names, mechanism of action, pharmacology, ADME, genetic variants associated with the drug efficacy or adverse drug reactions, related genes, key (hand-annotated) publication links, and also a VEGF-signaling map with targets for bevacizumab as well as other drugs targeting molecules in the pathway. Finally, under downloads/LinkOut tab are Wikipedia, DrugBank, PubMed, PubChem, etc., links. This database is a "cool" resource. See the VEGF pathway picture below. [cite a, b]
PharmGKB: a logical home for knowledge relating genotype to drug response phenotype. Altman RB. Nat Genet. 2007 Apr;39(4):426. | PubMed | Scholar |