Personalized medicine stands or falls with reliable laboratory data

From DNA analysis to medication advice: the quality of laboratory data helps determine how reliably personalized medicine can be applied in practice.

Personalized medicine promises to better tailor treatments to the individual patient. Genetic information can help with this. For example, some patients metabolize a drug faster than others, meaning that the same dosage may have insufficient effect in one patient and actually cause more side effects in another.[2]

However, there are several steps between DNA and a treatment recommendation. The first is perhaps the most fundamental: the reliable identification of genetic variants.

Therefore, the seminar takes center stage during World of Laboratory 2026. Personalized Medicine in Practice: From Data to Decisions on the program. It centers on the step from genetic and biological data to an substantiated clinical decision. Pharmacogenomics is an important application in this regard.

From DNA to medication advice

Pharmacogenomics investigates how genetic differences influence the action and breakdown of medicines. Genetic factors provide an important explanation for some of the differences between patients regarding the efficacy and side effects of medicines.[2]

In recent years, guidelines have been developed to translate that genetic information into practice. Organizations such as the Clinical Pharmacogenetics Implementation Consortium (CPIC) and the Dutch Pharmacogenetics Working Group (DPWG) link specific genetic variants to treatment and dosing recommendations for medicines.[3]

As a result, pharmacogenomics is increasingly becoming part of clinical practice. However, before a doctor or pharmacist can use such advice, it must first be reliably determined which genetic variants are present in a patient.

That is where the work of the laboratory begins.

Not every DNA test sees the same thing

In practice, there is no single standard method for genetic analysis. The choice of technology depends on the question the laboratory needs to answer.

A relatively targeted approach is panel genotyping. This involves looking for pre-selected genetic variants known to be relevant to drug response. Such an approach can be efficient, standardized, and relatively cost-effective.[3]

The limitation is clear: a panel looks for variants that have been pre-selected. Variants that are not included in the panel are not found. This can be relevant when new variants are discovered or when genetic diversity between populations starts to play a greater role.[3]

Another possibility is sequencing. With next-generation sequencing, much larger parts of the DNA can be examined, from individual genes to exomes and complete genomes. This provides a more comprehensive picture of genetic variation. At the same time, the amount of data that needs to be analyzed and interpreted is growing.[2]

More data therefore does not automatically mean more certainty. The quality of the analysis, bioinformatics, and interpretation becomes at least as important.

The right technology for the right question

That makes the choice of analysis technology increasingly important.

A targeted genotyping panel may be suitable when it is clear in advance which variants to look for. Short-read sequencing can yield much more genetic information, but has limitations in certain complex regions of the genome. Long-read sequencing, on the other hand, can offer advantages in such regions.[4][5]

Many pharmacogenes have very similar DNA sequences, repeats, and structural variants. As a result, these genes are not always easy to fully characterize using traditional methods. Long-read sequencing makes it possible to read much longer DNA fragments at once and can therefore better map complex genetic structures.[5][6]

An example is CYP2D6, an important pharmacogene involved in the breakdown of various drugs. Due to duplications, deletions, and other structural variations, among other factors, it can be difficult to fully characterize this gene with shorter reads. Long-read sequencing can help to better identify such variation and the combination of variants – the haplotype.[4][7]

That does not mean that long-read sequencing will replace all other methods. Cost, speed, throughput, validation and clinical implementation all play a role.[4][5]

The relevant question is therefore not which technology is the most advanced, but:

Which technology best suits the clinical question?

Data alone is not enough

Even when a laboratory has reliably identified a genetic variant, that is only the beginning of the next step.

A variant only gains clinical significance when it is clear what it means for, for example, the activity of an enzyme or transport protein. This information must then be translated into a treatment recommendation.[3]

This requires knowledge bases, guidelines and software. The results must also be integrated into the clinical workflow. The implementation of pharmacogenomics therefore requires collaboration between laboratories, pharmacies, bioinformaticians, ICT specialists and prescribers, among others.[3]

The challenge thus shifts from producing data to giving meaning to data.

So the question is not just how much genetic information a laboratory can collect, but also whether that information is reliable, interpretable and usable when a treatment decision needs to be made.[2][3]

Does it also work in practice?

The question, of course, is whether pharmacogenomics subsequently actually leads to better care.

An important study is the European PREPARE study. In this multicenter study, 6,944 patients were examined in 18 hospitals, nine primary health centers and 28 public pharmacies in seven European countries. The participants received a pharmacogenetic panel of twelve genes or standard care.[1]

In patients with a clinically relevant gene-drug interaction, a clinically relevant adverse event occurred in 21.0 percent of patients in the genotype-controlled group, compared to 27.7 percent in the control group. For all participants, the percentages were 21.5 and 28.6 percent, respectively. The researchers reported an odds ratio of 0.70.[1]

The results show that genotype-driven prescribing can reduce the number of clinically relevant side effects.[1] At the same time, it is important not to exaggerate the results. The study primarily shows that a pharmacogenetic approach is feasible and can yield clinical benefits when genetic information is actually used in prescribing.

Critical reactions also appeared regarding the interpretation of the PREPARE results. A commentary in The Lancet stated that the benefits of pharmacogenetic testing are not entirely clear based on the study.[8]

It is precisely this discussion that underscores the importance of proper implementation and reliable diagnostics. A genetic test in itself does not change the treatment. Its value only emerges when the result is correctly interpreted and actually becomes part of clinical decision-making.

From data to decisions

This brings the core of personalized medicine into view.

Behind a medication recommendation, an entire chain of technology and expertise can lie: sample processing, DNA isolation, genotyping or sequencing, quality control, bioinformatics, interpretation, and ultimately the translation into a clinical decision.

Every step can affect the reliability of the final result.

For laboratories, the development of personalized medicine therefore means more than access to increasingly advanced sequencing technology. It also means thinking about validation, quality assurance, data analysis and how results are used in healthcare.[3][5]

The development of long-read sequencing clearly demonstrates this. The technology can better map complex pharmacogenes, but at the same time raises new questions regarding costs, infrastructure, data analysis and implementation.[5]

The most important question is therefore ultimately not how much DNA we can read.

The question is which information is reliable enough to base a treatment decision on.

That is where the laboratory plays an important role. And it is precisely there that personalized medicine converges with the technology, knowledge, and expertise needed to actually convert data into decisions.

Sources

[1] Swen JJ, et al. A 12-gene pharmacogenetic panel to prevent adverse drug reactions: an open-label, multicenter, controlled, cluster-randomized crossover implementation study. The Lancet. 2023;401:347–356.

[2] Pirmohamed M. Pharmacogenomics: current status and future perspectives. Nature Reviews Genetics. 2023;24:350–362.

[3] Kabbani D, et al. Pharmacogenomics in practice: a review and implementation guide. Frontiers in Pharmacology. 2023;14:1189976.

[4] Neu M, Yang Y, Scott SA. Long-read Sequencing for Germline Pharmacogenomic Testing. Advances in Molecular Pathology. 2023.

[5] Tafazoli A, et al. Leveraging long-read sequencing technologies for pharmacogenomic testing: applications, analytical strategies, challenges, and future perspectives. Frontiers in Genetics. 2025;16:1435416.

[6] Logsdon GA, Vollger MR, Eichler EE. Long-read human genome sequencing and its applications. Nature Reviews Genetics. 2020;21:597–614.

[7] Van der Lee M, et al. Application of long-read sequencing to elucidate complex pharmacogenomic regions: a proof of principle. The Pharmacogenomics Journal. 2021;22:75–81.

[8] The PREPARE study: benefits of pharmacogenetic testing are unclear. The Lancet. 2023.