Pharmacogenomics (PGx) landscapes and related databases are important for identifying biomarkers of drug response and toxicity. Beginning in 2001, PharmGKB was established for PGx investigations. Subsequently, the eMERGE-PGx network and the Sequence, Phenotype, & pHarmacogenomics INtegration eXchange (SPHINX) database were developed. Based on these and other databases, PGx guidelines [including the CPIC and Dutch pharmacogenetics working group (DPWG)], were developed hopefully to guide precision medicine in clinical studies. Recent studies validated that personalized drug-treatment-guided PGx panel tests appears to provide safer therapies for patients.
Because the genetic backgrounds of different populations are diverse, different countries have conducted PGx studies and projects on their own populations. For example, PGx projects and clinical implementation projects — such as U-PGx in Europe, National Laboratory for the Genetics of Israeli Populations (NLGIP) in Israel, and Southeast Asian Pharmacogenomics Research Network (SEAPharm) in Southeast Asia — were established.
However, a PGx landscape and database for Chinese individuals has still been lacking, although some large-scale sequencing projects [including the China Metabolic Analytics Project (ChinaMap), the Chinese Millionome Database (CMDB), the NyuWa project, the Stroke Omics (STROMICS) genome study, the HuaBiao project, and the Westlake BioBank for Chinese (WBBC) project] have been set up. A previous study by these authors on 28 variants in 22 actionable PGx pharmacogenes using real-world PGx test results from 12,578 patients also revealed that the pattern of pharmacogenes in the Chinese population are different from that in other populations. All these results indicated that data from other races can serve as a reference, but they should not be directly applied. Thus, these types of data are still lacking for the nationwide Chinese populations and are badly needed.
One of the most important aims of this study was to identify novel variants specific to the Chinese population. Authors concluded that a total of 331,610 novel variants in 3,538 “pharmacogenes”(!!) were identified — which could provide a large amount of supplementary PGx information for the Chinese population — similar to the projects conducted for other populations mentioned above (including eMERGE-PGx and TOPMed). Among all reported mutations, functional and novel genetic variants are most likely to have an impact on drug response and may be potential PGx biomarkers for the Chinese population. These Chinese-specific variants of metabolic enzymes, transporters, and receptors are listed in Supplementary Tables S8, S9, and S10, respectively. Therefore, a genetic test panel based on these variants may benefit Chinese patients.
Another aim of this study was to estimate the ethicity and geographical differences in pharmacogene variants in the Chinese population. A comparison of all mutation frequencies among different ethnic groups revealed that the genetic variant frequencies in the Chinese population were different from those in the European and African populations. For geographical and regional differences, authors compared the data from 31 provinces in China and found that the identified pharmacogene variants also differed substantially within China. Differences in pharmacogenetic variants between northern and southern China were also identified, and drugs potentially affected by these variants were systematically analyzed. All mutations with ethnicity and regional differences in this study are provided in the Supplemental Data tables, which can be accessed by anyone.
The authors’ study and the constructed database have three potential applications. First, the novel variants identified in this study can be used to discover new PGx biomarkers with increased efficiency. To validate this possibility, authors tested the potential effects of 26 novel mutations on platinum resistance in 1,019 patients with lung cancer, and 14 high-quality mutations were used for downstream analyses. Two novel variants were ultimately associated with the likelihood of a platinum drug outcome. Notably, numerous novel variants have been reported and may be identified for a large proportion of drugs, including future marketed drugs, as was done for platinum in this study — without conducting additional whole-genome sequencing (WGS) studies.
Second, individual differences and PGx biomarkers for new drugs can be rapidly evaluated using these data. This information would be especially important for the treatment of sudden outbreaks of diseases — such as Cov-Sars-2 (COVID19) infections. In one of the previous studies by these authors, genetic data from Asian and American populations were used to identify genetic variants that may affect the efficacy of remdesivir during the COVID19 pandemic. These mutations may explain the ethnic differences in drug responses described between the United States and China. Similarly, investigators can rapidly predict whether there are individual and racial differences in future responses with other new drugs.
Third, the genetic variants identified in this study can be used to design Chinese-specific PGx test panels, which can be used to provide personalized medicine guidance for drug utilization.
The current study has several limitations. First, the low-depth sequencing data generated from non-invasive prenatal testing (NIPT) were used for PGx analyses in this study — making it difficult to accurately identify copy number and structural variants. High-depth and long-read sequencing technologies were used in phase II of this study. The novel variants discovered in this study should therefore be validated in a larger Chinese cohort in the future.
Second, although the three main classes of pharmacogenes were analyzed systematically, a deeper interpretation of other pharmacogenes and some low-frequency variants should also be considered. The clinical efficacy or safety of drugs related to these PGx variants in the Chinese population need to be further validated. The biological validation and the potential mechanism underlying these variants are important for their effective application in clinical practice.
Third, the demographic structure of the Chinese population is complex. The participants were mainly from East and Central China, and most were from the Han ethnic group. Participants from other regions and ethnic groups should be increased in number — to improve the PGx landscape in China.
Nevertheless, in total, 1,577,513 variants (including 331,610 novel variants) were identified among 3,538 pharmacogenes related to 2,086 drugs. The average numbers of drug enzyme, transporter, and receptor variants were 258, 557, and 632, respectively. Subsequent correlation analysis indicated that those pharmacogenes affecting multiple drugs had fewer variants. Among the 16 categories of drugs, authors found that nervous system, cardiovascular system, and genitourinary system/sex hormone drugs were more likely to be affected by variants of pharmacogenes. Characteristics of the variants in the metabolizing enzyme, transporter, and receptor subfamilies showed specificity. Finally, a pharmacogenomic database, named the Chinese Pharmacogenomics Knowledge Base (CNPKB: http://www.cnpkb.com.cn/
CNPKB（中国人群药物基因组学知识库）
No. 110 Xiangya Road, Changsha, Hunan Province, China Email: CN-PKB@csu.edu.cn. 湘 ICP 备2023018707
www.cnpkb.com.cn
), was constructed for including all the data. Authors believe that this study (which constructed a PGx landscape and database for a large Chinese cohort) can serve as a guide for personalized Chinese drug treatment in the future.
My colleague Magnus Ingelman-Sundberg (Stockholm) noted that “these authors analyzed cell-free DNA (cfDNA) from plasma — instead of genomic DNA (gDNA). And cfDNA is notoriously fragmented or under-represented, compared to gDNA. In the Figure 2 panels, nearly half the identified variants have a frequency greater than 1%, with the vast majority of variants being intronic. In addition, most of the exon variants identified were located in untranslated regions (UTRs). Their claim of identifying 3,538 “pharmacogenes” is highly questionable, and their detection of e.g., only 24 single-nucleotide variants (SNVs) in the CYP2D6 gene raises further serious concerns. Overall, I find the data are unreliable.” I hope to hear also from Professor Ge Zhang?
DwN
COMMENT:
If you could respond to Dr. Nebert’s email blog, especially the comment on your data quality (unreliable) in the last paragraph, that would be much appreciated.
Sincerely,
COMMENT:
Dear Professor Zhang:
Thank you very much for you and your colleagues’ interest in our study. We are very happy to discuss these concerns. We have written a long document to explain all questions; please see attached file. To improve our study, we are also willing to discuss further these and any other concerns.
COMMENT:
Dear Dr. Yin,
Thank you very much for your prompt reply and the detailed explanation. I copied your message to Dr. Nebert, who has a large mailing list which includes researchers in this field of study.
Dear Dr. Nebert
For the attached pdf file, please see the reply from Dr. Yin, who is one of the two corresponding authors. I am interested in this study, but I have limited expertise. I hope you can share Dr. Yin’s message in your email blog list and please copy both Drs. Yin and Wang. I look forward to hearing from you.
Thank you All,
COMMENT:
Dr Xiang Zhang (Professor and Director of the Genomics, Epigenomics and Sequencing Core, at University of Cincinnati) took it upon himself to contact by email the two corresponding authors of this herculean study published 2 Jan 2025 [first attachment] and to ask them to respond to the questions listed by Professor Magnus Ingelman-Sundberg in the last paragraph. Attached [second attachment] is the coauthors’ (very thorough) response to Professor Ingelman-Sundberg’s query.
—DwN