If you cannot find the information that you are looking for, then you can request assistance from the PharmGKB team.
The Pharmacogenomics Knowledge Base, PharmGKB, is an interactive tool for researchers investigating how genetic variation effects drug response. A review of 10 years of PharmGKB can be found in "Pharmacogenomics and bioinformatics: PharmGKB" [Article:20350130].
The actual data reside on secured computers at the Stanford University campus.
The United States National Institutes of Health (NIH) funded both the researchers that collected the data, as well as the researchers who are developing PharmGKB. These include the following NIH components:
You can apply to become a registered user by clicking the "SIGN IN" link at the top of the website next to the search box. Then follow the "Have you registered?" link. You will need to fill out an application form stating your reason for wanting to register and your contact information. You will receive an email with a link to follow to activate your account.
After signing in with the current password, click the "My PharmGKB" tab at the top of the web page. Click the second link listed on that page – "Change Password". Type in your current password and the new password (two times to confirm).
You are welcome to link to PharmGKB but you may not redistribute any data/information from PharmGKB. If you would like to link to PharmGKB please use the following formats for URLs. The PA numbers can be obtained from the Downloads section.
For genes: http://www.pharmgkb.org/gene/PAxxx
For drugs: http://www.pharmgkb.org/drug/PAxxx
For diseases: http://www.pharmgkb.org/disease/PAxxx
For RSID: http://www.pharmgkb.org/rsid/rsxxx
In citing the PharmGKB please refer to:
M. Whirl-Carrillo, E.M. McDonagh, J. M. Hebert, L. Gong, K. Sangkuhl, C.F. Thorn, R.B. Altman and T.E. Klein. "Pharmacogenomics Knowledge for Personalized Medicine" Clinical Pharmacology & Therapeutics (2012) 92(4): 414-417.
If you are citing a pathway please check the pathway diagram page to find out whether the pathway has been published in the Pharmacogenetics and Genomics journal. If there is such a citation on the pathway page, please use that. If there is not, please use the PharmGKB citation for the publication titled "Pharmacogenomics Knowledge for Personalized Medicine".
Also, please send a brief email to email@example.com to inform us of which data you are using and for what purpose.
No, PharmGKB does not sell any drugs or recommend suppliers of drugs.
There are many databases that have some of the information in PharmGKB, such as PubMed, Drug Bank and dbSNP, and we rely on their resources. PharmGKB brings together the relevant data in a single place and adds value by combining disparate data on the same relationship, making it easier to search and easier to view the key aspects and by interpreting the data. We provide clinical interpretations of this data, curated pathways and VIP summaries which are not found elsewhere.
PharmGKB has data about genes, drugs and diseases and how these are interrelated. Data has been curated - extracted and sorted - from multiple sources to bring together information that builds a broader picture of the knowledge surrounding these relationships in a searchable or computable manner. The knowledge is presented as genomic variant annotations (that display the key data from individual publications), clinical annotations (that summarize several publications and interpret the data by genotype), VIP gene summaries (that give textual information about Very Important Pharmacogenes), drug centered pathways (in pictorial, textual and computer readable formats) and literature annotations (genes, drugs and diseases discussed in publications). See more information on our Overview page.
Details of our policies for data usage can be found on our Policies page.
Data can be viewed using the website or downloaded from links under the Download tab on the dark blue toolbar at the top of every PharmGKB webpage. A user account and agreement to the PharmGKB database license agreement is necessary for downloading data. This can be done in bulk using web services or as zipped up packages of spreadsheets with literature annotations, variant annotations, clinical annotations or pathway relationships.
See the Data Usage Agreement.
The PharmGKB is run by a multidisciplinary team that brings together computer scientists, software engineers, developers and research scientists with advanced degrees in the biological sciences (referred to as curators at PharmGKB) to construct frameworks and displays to curate pharmacogenomic data.
No. PharmGKB manually curates articles on a routine basis from the major pharmacogenomics journals and individual articles found in the process of compiling literature reviews for VIPs and pathways. This corpus of articles does not represent the entire pharmacogenomics literature in PubMed. PharmGKB is developing methods to automatically identify pharmacogenomics literature in PubMed and the results of these are available in the "Non-Curated Information" section at the bottom of the pages of the "Related To" tab.
Please contact us using the feedback button on the page with the error and describe the problem. Please make sure that you specify the page (URL, if possible) that has the error, and how you arrived at that page (eg. internal PharmGKB search or from an external website). Thank you for taking the time to do this and improving the resource for all users.
Clinical annotations, genotype-based dosing guidelines, drug labels, genetic tests, variant annotations, haplotypes, VIP summaries, drug pathways, literature annotations and pharmacogenetics summaries for top drugs (see "Drugs in PharmGKB" section) are manually curated in PharmGKB.
Please contact us using the feedback button and give the details of the publication you would like us to consider annotating. Thank you for taking the time to do this and improving the resource for all users.
If there is a PA number in the paper do a search on that. Alternatively, do a keyword search for a gene or drug studied in the paper and look at the datasets listed under the Downloads/Linkout tab. Or if the data are from a consortium there is a link to the submissions on the right panel of the consortium page (found by doing a keyword search for the consortium). If you still don't find what you are looking for please contact feedback for assistance.
Literature annotations and variant annotations are added in real time as a curator annotates the paper, so the tables on the "PGx Research" and "Is related to" tabs can change during the course of a day. VIPs, pathways and clinical annotations are reviewed periodically (VIPs and pathways have about a 2 year cycle, shorter if new evidence is readily apparent, clinical annotations are reviewed annually). New guidelines, drug labels and tests are added as they become evident.
Gene, drug and disease data that is automatically retrieved is updated in response to scheduled updates at the external resource.
PharmGKB collects some frequency data for genomic variants involved in pharmacogenomic relationships although it is not a primary focus and other resources may be better for finding extensive information. The frequency data of variants within individual studies is noted in the population section of variant annotations where available. This frequency data is taken directly from the annotated article. The primary datasets submitted to PharmGKB by the PGRN also contain some frequency data. Some of the VIP variant summaries also have frequency tables listing data from various publications.
Other resources that have frequency data include dbSNP, ALFRED and HapMap.
While PharmGKB does not routinely curate genomic variants associated with disease risk, some publications encountered during routine work may have these associations and they may be entered into the PharmGKB at the curators' discretion.
No, curating genomic variant data is a highly time consuming process. While PharmGKB strives to routinely curate those associations reported in the major pharmacogenomics journals as well as historically important studies encountered during literature reviews for VIPs and pathways, we acknowledge that we cannot manually curate the literature in its entirety. If you see something you think we should have, please send us feedback.
Genetic variants that are generally associated with drug response or outcomes are tagged with PD, those specifically associated with ADME of a drug are tagged with PK.
PharmGKB curators record the reported race and ethnicity in published studies when creating a variant annotation. PharmGKB uses OMB-defined categories (found here). If a study population contains people from multiple race/ethnicity groups, the study is labeled "mixed population". If the study population is not declared in the article, the study is labeled "unknown population".
A variant annotation reports the association between a single variant (single-nucleotide polymorphism or haplotype) and a drug phenotype from a single publication. These annotations are created manually by Scientific Curators who read each paper and extract the key information, including relevant drugs, study size, population data, statistical values, etc. and map the variants to a common standard (dbSNP rs#). Multiple variant annotations may be created for a single publication if it reports multiple associations between variants and drugs. Likewise, multiple variant annotations may be created for a single variant if multiple publications reporting an association between that variant and drugs have been curated by PharmGKB curators.
It is important to understand that the PharmGKB Scientific Curators routinely review several high profile journals for articles to curate. However, there may be more literature in the public domain to support or contradict a pharmacogenetic association that has not been curated by PharmGKB db. PharmGKB does its best to manually curate high profile literature but does not contain curated literature from every domain-based journal, or all of PubMed.
For more information about variant annotations, see McDonagh EM, Whirl-Carrillo M, Garten Y, Altman RB, Klein TE (2011). From pharmacogenomic knowledge acquisition to clinical applications: the PharmGKB as a clinical pharmacogenomic biomarker resource. Biomarkers in Medicine, 5(6):795-806 [Article:22103613].
Clinical annotations build on variant annotations, often combining multiple variant annotations into a single summary of the relevant variant-drug-phenotype association. The genomic variant annotations, each from a single publication, are manually examined and the overall relevance described for each possible gentoype. Essentially, a clinical annotation is a genotype-based summary of the clinical impact of a genomic variant based on the information at PharmGKB.
The phenotype for any given genotype is reported in a relative fashion as compared with other genotypes. For example, the AA genotype for a variant may be associated with an increased risk of side effects as compared with the AG and GG genotypes - but not necessarily at an increased risk of side effects for patients on the drug in general, as this would depend on a detailed examination of the target population allele frequencies and the populations on which the original US Food and Drug Administration approval is based.
The annotation is scored for the strength of evidence used to compile the annotation on a scale of 1A to 4 with 1A being the strongest evidence. For information about how the strength of evidence is assigned, see our Clinical Annotation Levels of Evidence page.
For more information about clinical annotations, see:
Whirl-Carrillo M, McDonagh EM, Hebert JM, Gong L, Sangkuhl K, Thorn CF, Altman RB, Klein TE (2012). Pharmacogenomics Knowledge for Personalized Medicine. Clinical Pharmacology and Therapeutics, 92(4): 414-7 [Article:22992668].
McDonagh EM, Whirl-Carrillo M, Garten Y, Altman RB, Klein TE (2011). From pharmacogenomic knowledge acquisition to clinical applications: the PharmGKB as a clinical pharmacogenomic biomarker resource. Biomarkers in Medicine, 5(6):795-806 [Article:22103613].
See our Clinical Annotation Levels of Evidence page.
For more information, see Whirl-Carrillo M, McDonagh EM, Hebert JM, Gong L, Sangkuhl K, Thorn CF, Altman RB, Klein TE (2012). Pharmacogenomics Knowledge for Personalized Medicine. Clinical Pharmacology and Therapeutics, 92(4): 414-7 [Article:22992668].
A literature annotation is the tagging of a single publication with the genes, drugs and diseases discussed in the article. A literature annotation does not necessarily imply a direct relationship between the genes, drugs and diseases discussed.
PharmGKB uses the term publication to refer to a published article that has been catalogued in the PubMed database.
VIP stands for Very Important Pharmacogene. VIP summaries provide an overview of a significant gene involved in metabolism of or response to one or several drugs. Often, VIPs either (1) play a role in the metabolism of many drugs (eg. CYP2D6); or (2) contain variants which potentially contribute to a severe drug response (eg. HLA-B). VIP summaries typically include background information on the gene including any disease associations, as well as in-depth information on the gene’s pharmacogenetics. VIP genes are chosen through extensive review of a variety of sources, including the U.S. Food and Drug Administration (FDA) biomarker list, FDA-approved drug labels with pharmacogenetic information, and Clinical Pharmacogenetic Implementation Consortium (CPIC) nominations. Additionally, VIPs may be created if a gene is associated with a large number of variant annotations and is part of high-level clinical annotations. VIP summaries are often published in the journal Pharmacogenetics and genomics.
PharmGKB has a wish-list of VIP genes to develop, based on a survey of the PGRN. If you would like to work with PharmGKB to develop a VIP summary please contact feedback.
PharmGKB pathways are evidence-based diagrams depicting the pharmacokinetics (PK) and/or pharmacodynamics (PD) of a drug with relevant (or potential) pharmacogenetic (PGx) associations. Drugs featured in PharmGKB pathways are chosen through extensive review of a variety of sources, including, but not limited to, the U.S. Food and Drug Administration (FDA) biomarker list and Clinical Pharmacogenetics Implementation Consortium (CPIC) nominations.
PharmGKB pathways are accompanied by a written summary of the PK and/or PD pathway, as well as other important PGx related information. Interactions within each pathway are supported by manually curated evidence from published literature and specific information for each arrow in the pathway is found on the “Components” tab. Information contained within each pathway is available for download in TSV, BioPAX and GPML formats. The pathway graphic is available for download in PDF file format. PharmGKB pathways are often published in the journal Pharmacogenetics and Genomics.
If you would like to work with PharmGKB to develop a drug pathway please contact feedback.
Yes, please acknowledge the copyright to PharmGKB and state that permission has been given by PharmGKB and Stanford University. In addition, please check the pathway diagram page to find out whether the pathway has been published in the Pharmacogenetics and Genomics journal. If there is such a citation on the pathway page, please use that. If not, please use the PharmGKB citation.
Also, please send a brief email to firstname.lastname@example.org to inform us of which pathway diagram you are using and for what purpose.
The dosing guidelines are suggested modulations to the existing dosing guidelines that incorporate genotype based recommendations that have been published by the Clinical Pharmacogenetics Implementation Consortium (CPIC) or the Royal Dutch Association for the Advancement of Pharmacy - Pharmacogenetics Working Group (DPWG).
The drug label pages have the FDA drug label for drugs that have pharmacogenomic information and the relevant sections are manually highlighted by PharmGKB. The list of drugs with label information is derived primarily from the information found at the "Table of Pharmacogenomic Biomarkers in Drug Labels". PharmGKB does NOT contain all drug labels, but PharmGKB drug pages link to DailyMed in cases when that information is available to us.
The term haplotype refers to a cluster of allelic variants (including SNPs, insertions, deletions, etc) that are inherited together as a consequence of their proximity to one another on the chromosome. PharmGKB curates haplotype to allelic variant translation tables for certain genes where the information is available. If a gene has a haplotype translation table available it can be found on a tab labeled “Haplotype” on the gene page. If there is no “Haplotype” tab on a gene page then the gene does not have a haplotype table available. PharmGKB does not define haplotypes. Information on how PharmGKB sources haplotype translation tables can be found below in the section "Questions about data sources."
Yes PharmGKB still has primary data although it is no longer our primary focus. Primary data can be found in the datasets section of the Downloads/LinkOuts tab on a drug, gene or disease page, or the "Download" tab in the blue toolbar located on every PharmGKB page.
If you wish to submit primary data please contact the curators at feedback.
Yes, allele frequencies on PharmGKB come from two sources: (1) an article that is annotated and (2) HapMap phases II+III. Allele frequencies from articles exist as part of Variant Annotations. Allele frequencies displayed on the "Overview" tab of Variant pages are calculated from HapMap phases II+III.
Here is a list of data sources that PharmGKB uses and their versions.
PharmGKB uses dbSNP rs identifiers to map genomic variants. This mapping is often done manually by PharmGKB curators. We attempt to capture the multiple text-based names used in the literature as we come across them, so while we have common name information, these lists are not exhaustive. We also track how variants map to the UCSC Golden Path genome map and the genomic build from Entrez.
Allele frequencies on PharmGKB come from two sources: (1) an article that is annotated and (2) HapMap phases II+III. Allele frequencies from articles exist as part of Variant Annotations. Allele frequencies displayed on the "Overview" tab of Variant pages are calculated from HapMap phases II+III.
PharmGKB collects information on haplotype definitions from different sources. For example, the haplotypes that define cytochrome P450 (CYP) alleles are derived from the The Human Cytochrome P450 (CYP) Allele Nomenclature Database, and haplotypes that define UDP glucuronosyltransferase (UGT) alleles are derived from the UGT Alleles Nomenclature page, which are maintained by the Karolinska Institute and the Université de Laval, respectively.
In many cases, however, haplotypes for a particular gene are reported and defined by individual studies. There is no centralized resource or entity responsible for reconciling haplotypes or haplotype names published by different authors for these genes. Individual authors may or may not acknowledge previous haplotype publications. They may or may not interrogate the same positions or follow the same naming conventions as in previous publications. This can result in discordance in the information from different publications.
In these cases, PharmGKB attempts to collect the information from these studies into one location. PharmGKB does not attempt to standardize haplotype definitions or names. If a study provides the sequence changes that define a haplotype but does not provide a name for the given haplotype, PharmGKB may use sequential numbering to provide each haplotype with a name. This is not an “official” haplotype name, but rather a designation PharmGKB creates in order to distinguish amongst haplotypes.
Descriptions on how specific haplotype sheets are sourced and named appear on the haplotype tab for each gene with a haplotype sheet. If haplotypes are sourced from individual studies, the sources for each haplotype appear under the column “PMID".
PharmGKB imports the drug names and associated metadata (structures, properties, indications) mostly from DrugBank citations. Information from DrugBank is not routinely manually curated by PharmGKB and is labeled with the tag "Source:DrugBank" at the end of each section (however, if you find an error let us know and we will correct it and inform DrugBank of the problem). Additional drugs not found in DrugBank are occasionally imported from PubChem or other sources and are labeled as such.
PharmGKB imports the standardized disease vocabulary found in MeSH: Medical Subject Headings or SnoMED. The overview page shows the inter-related MeSH, SnoMED and UMLS terms.
Known Caveats: While we recognize that the term "disease" does not accurately describe some of the phenotypes in MeSH (eg. Death) and that there are several terms that are hierarchically related, we do not have an alternative freely available structured ontology for diseases/phenotypes at present. If you wish to assist PharmGKB in this matter please contact feedback.
PharmGKB does not routinely annotate variant-disease associations, including causal or associated SNPs. Most disease annotations are tangentially related to variant-drug associations, eg. the study population may be disease-based but the drug discussed may or may not be used for that indication and the gene/variant is typically associated with the drug, not the disease.
In general pharmacogenetics usually refers to how variation in one single gene influences the response to a single drug. Pharmacogenomics is a broader term, which studies how all of the genes (the genome) can influence responses to drugs. However, these terms are often used interchangeably.
A pharmacogenetic study is a subset of genetic studies. The tests in a pharmacogenetic study attempt to predict how a person will respond to a drug. It might be a test that has nothing to do with a disease at all, but about how quickly the body breaks down the drug so the person's doctor could give a different dose if the person is likely to break down the drug very slowly or very fast. Or the test result might suggest that the person has a risk of a bad reaction to a drug so the doctor should use a different drug. But sometimes it might be a test that relates to the person's disease profile. E.g. for the cancer drug Herceptin, a drug that works on tumors that have a particular biomarker,the pharmacogenetic test is performed to assess if the tumor will be attacked by the Herceptin drug.
The same as with genetic tests to determine risk of a disease, the results of pharmacogenetics tests are not deterministic. Even if a person's genetic test shows they have a gene variant associated with an increased risk for heart disease, they may not develop heart disease. Even if the pharmacogenetic test says a person doesn't have the marker for high risk for a bad response to codeine, they might still have a bad response to codeine due to some other factor. Multiple genetic and environmental factors can contribute to an individual's response to a drug, as well as interactions with drugs taken concomitantly.
Personalized medicine is using information specific to a patient to tailor their health care. Pharmacogenomics can be used as part of a personalized medicine approach.
There are currently many obstacles to the use of pharmacogenomics in routine patient care. These include but are not exclusive to: lack of evidence that pharmacogenomics guided therapy is better than regular therapy; cost or lack of insurance coverage of pharmacogenetic testing; physicians not feeling sufficiently trained to apply results of pharmacogenetic tests; lack of guidelines in how to interpret test results; lack of studies showing how to develop pharmacogenomic associations into predictive tests. This is discussed further in these articles - PMIDs: 21326263, 21884816, 21691271, 21619429, 20485318.
PK is short for Pharmacokinetics, the movement and change of drugs in the body over a period of time. Genetic variation in processes involved in the absorption, distribution, metabolism, or elimination of a drug can result in changes in drug availability. Data in this category are primarily concerned with demonstrating that genetic polymorphisms lead to variations in the levels or concentrations of drugs or their metabolites at the site of action.
PD is short for Pharmacodynamics and Drug Response, the study of the biochemical and physiological effects of drugs and the mechanisms of their actions. Genetic variation in drug targets can cause measurable differences in the response of an organism to a drug. Data in this category document that the biological or physiological response to a drug varies, and that this variation can be associated with the variation of one or more genes. This variation is often measured at the whole-organism level.