Pharmacogenomic landscape of a Moroccan cohort: enhancing global diversity and addressing the North African gap Authors: Souad Kartti, Mohammed Walid Chemao-Elfihri, Mohammed Hakmi, Marouane Aherkou, Rachid ElJaoudi, Lahcen Balouch, Elmostafa El Fahime, Lahcen Belyamani, Saber Boutayeb Published: 2026 ABSTRACT AbstractPharmacogenomics (PGx) enables personalized therapy by considering individual genetic variation. However, North African populations remain underrepresented in global PGx databases despite their substantial and unique genetic diversity, which can strongly influence drug response. This study characterizes the pharmacogenomic landscape of 109 Moroccan individuals using whole-genome sequencing data. Across 8289 candidate pharmacogenes, we identified 129,021 novel variants, predominantly within DPYD, CYP2C19, and SLCO1B1, underscoring the region’s considerable genomic diversity. Analysis of 17 clinically relevant pharmacogenes with established CPIC drug guidelines revealed the high prevalence of SLCO1B1*15, UGT1A1*80 + *28, CYP2C19*17, and VKORC1 rs9923231 alleles. Additionally, HLA-A and HLA-B genotyping identified frequent alleles (HLA-A*03:01, HLA-A*30:02, HLA-B*51:01) previously associated with adverse drug reactions. Collectively, these findings provide a foundational resource for population-specific pharmacogenomic interpretation and support the advancement of precision medicine in Morocco and across North Africa.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-026-54350-1.Keywords: Pharmacogenomics, WGS, Precision medicine, Drug, CPIC, ADRSubject terms: Computational biology and bioinformatics, Diseases, Genetics, Medical research Introduction Pharmacogenomics, which aims to customize drug therapies based on an individual’s genetic makeup, plays a central role in the evolution of precision medicine by enhancing therapeutic efficacy and minimizing adverse drug reactions (ADRs)1,2. Numerous studies have identified substantial interethnic disparities in allele and genotype frequencies of genes encoding drug-metabolizing enzymes, transporters, and targets. These differences can significantly alter pharmacokinetics and pharmacodynamics, and increase the risk of adverse drug reactions. Despite significant progress in recent years, most large-scale pharmacogenomic studies and databases still largely reflect populations of European or East Asian ancestry3,4 . This imbalance poses a challenge for the equitable implementation of precision medicine worldwide. Numerous studies have shown that the frequencies of pharmacogenetically relevant variants vary widely across populations, sometimes leading to important differences in drug efficacy and safety profiles5,6 . In Australia, a recent pharmacogenomic study predicted that over half of the Tiwi population have altered metabolism of CYP2C9, CYP2C19, and CYP2B6, affecting the response to common analgesics, statins, anticoagulants, antiretrovirals, antidepressants, and antipsychotics7. Additionally, a recent study involving 881 Serbian individuals revealed notable differences in the distribution of CYP2B6, NAT2, SLCO1B1, UGT1A1, and VKORC1 alleles compared to other Europeans, with direct implications for prescribing statins and antidepressants8. Similarly, in the Middle East, the pharmacogenomic analysis of more than 14,000 Qataris demonstrated a high prevalence of actionable metabolizer phenotypes for CYP2C19, CYP2D6, and CYP2B6, with major clinical consequences for psychiatric drug management9. African populations, and North Africans in particular, exhibit important genetic diversity that is underrepresented in global databases such as 1000 Genomes and gnomAD10,11. Notably, the Moroccan population, with its distinct genetic profile12 , may show unique distributions of alleles affecting drug metabolism and response, highlighting the importance of including these populations in pharmacogenomic research. Although pharmacogenomic data from North African populations remain limited, a study of pain-management variants in Mediterranean populations revealed regional differences in allele frequencies that could influence analgesic efficacy13. Similarly, several studies involving Moroccan individuals have contributed to early investigations of genotype–phenotype associations relevant to clinical practice. For instance, it has been reported that cytochrome P450 (CYP450) enzymes and ABCB1 polymorphisms influence clopidogrel responsiveness in Moroccan patients with acute coronary syndromes, highlighting variability in antiplatelet drug response14. In addition, another Moroccan study demonstrated that APOBEC3G variants, particularly rs2294367, modulate susceptibility to HIV-1 infection and treatment response, underscoring the broader relevance of host genetic factors in disease outcomes15. These findings reinforce the need for broader, population-level studies to comprehensively characterize pharmacogenomic diversity in Morocco. In this study, we present the first pharmacogenomic landscape of a cohort of 109 whole-genome sequenced Moroccan individuals12. We focused on variants within clinically actionable pharmacogenes, aiming to identify allele frequencies, genotype distributions, and predicted phenotypes that may impact drug response in this population. Our findings provide a foundational resource for precision healthcare initiatives in Morocco and help address the broader gap in pharmacogenomics data from North African populations. Materials and methods Cohort description and sequencing data The genomic data analyzed in this study were obtained from a cohort of 109 healthy Moroccan individuals, who provided written informed consent. The study protocol was approved by the relevant ethics committees. The genomic data analyzed in this study were obtained from 109 healthy Moroccan individuals. The study was approved by the Ethics Committee of the Faculty of Medicine and Pharmacy of Rabat, Mohammed V University, Rabat, Morocco (Approval No. 10/15), and all participants provided written informed consent. All methods were conducted in accordance with the relevant guidelines and regulations, including the Declaration of Helsinki. Whole-genome sequencing (WGS) was performed as part of the first phase of the Moroccan Genome Project (MGP), as detailed in our recent publication12, which serves as the primary reference for this dataset. Briefly, sequencing was conducted using the Illumina NovaSeq6000 platform, generating PCR-free 150 bp paired-end reads at an average depth of 30x. Sequence alignment to the GRCh38 human reference genome was carried out using the BWA-MEM algorithm16, and variant calling followed the Genome Analysis Toolkit (GATK) Best Practices pipeline17. For the purposes of the present study, analyses focused specifically on pharmacogenomic markers extracted from this previously described WGS dataset. Variants quality control The resulting multi-sample VCF file was normalized, and multi-allelic sites were split using BCFtools v1.15.118 to optimize the dataset for downstream analyses. Variants with more than 11 missing genotypes (approximately 10% of missingness) were then eliminated to ensure the quality and completeness of the data. Before Hardy–Weinberg equilibrium analysis, variants on the Y and mitochondrial chromosomes were excluded to avoid bias due to their sex-linked mode of inheritance. A p-value threshold of 5e−07 was applied to identify and filter variants showing significant deviations from Hardy–Weinberg equilibrium, using PLINK219 after data conversion to bed/bim/fam format. Variant annotation Genetic variant annotation was performed using the Ensembl Variant Effect Predictor (VEP) version 113.0 with the ClinVar 2025 (clinvar_20250217.vcf.gz) to assess functional consequences and clinical significance. The multi-sample VCF file served as input data, and default parameters were employed with the additional flag (–custom file = "./clinvar_20250217.vcf.gz") to prioritize ClinVar 2025 pathogenicity classifications and variant consequences. Drug-related variants selection A comprehensive catalog of pharmacogenomically relevant genes was compiled by integrating data from ChEMBL20, DrugBank21, and ClinPGx (Formerly PharmGKB)22 databases. Genes were selected based on their established roles in drug metabolism, transport, target interactions, and related pathways. Non-protein-coding genes and redundant entries were excluded from the final dataset. Variants within these genes were extracted from the annotated VCF and subsequently filtered to include only those with existing ClinPGx annotations, distinguishing between clinically actionable variants and those supported solely by functional evidence. Identification of potentially deleterious novel variants We prioritized novel variants using Combined Annotation Dependent Depletion (CADD) v1.723 , which integrates diverse genomic features into a single Phred-scaled score reflecting predicted functional impact and has shown strong performance in pharmacogenomic variant interpretation24. Variants with CADD Phred scores > 20, corresponding to the top 1% of deleterious variants genome-wide, were classified as potentially deleterious. Assessment of variants in clinically relevant pharmacogenes To establish a pharmacogenomic profile of the Moroccan population, we selected 17 clinically relevant pharmacogenes for which the Clinical Pharmacogenetics Implementation Consortium (CPIC) (https://cpicpgx.org/; accessed 24 July 2025) has published therapeutic guidelines related to different drug use. The selection was based on the CPIC classification system, including only gene/drug pairs with Level A annotations. Among the studied pharmacogenes, 11 (CYP2D6, CYP2C9, CYP2C19, CYP3A5, CYP4F2, CYP2B6, UGT1A1, SLCO1B1, TPMT, DPYD, and NUDT15) were investigated for star allele identification. These genes have standardized allele nomenclature officially recognized by specialized entities such as PharmVar (https://www.pharmvar.org/), the TPMT Nomenclature Committee (https://liu.se/en/research/tpmt-nomenclature-committee), or the UGT Nomenclature Committee (https://www.pharmacogenomics.pha.ulaval.ca/ugt-alleles-nomenclature/). The analysis of the remaining genes (VKORC1, IFNL3, ABCG2, RYR1, G6PD, and CACNA1S) focused on specific well-characterized variants known to affect drug response. Before allele calling, computational phasing was performed using SHAPEIT525 to assign variants to the paternal and maternal chromosomes. The resulting phased VCF file was then analyzed using PyPGx v0.2526, Stargazer v2.027, Aldy v4.728, and PharmCAT v2.15.529 for alleles and diplotypes assignment. For genes known to harbor structural variants, notably CYP2D6, CYP2B6, CYP4F2, and G6PD, the analysis was conducted using longv1.0.330, PyPGx, Stargazer, and Aldy, which are suitable for detecting complex variants. Depending on the tool, inputs included either CRAM files alone or combined with the VCF files. To select the final diplotypes, cases where all tools produced concordant results were straightforwardly retained. In contrast, when discrepancies arose between tools, manual inspection was performed to confirm the presence of the relevant variants. The CPIC diplotype/phenotype tables were used to assign the corresponding phenotype to each diplotype for each gene. HLA Genotyping and pharmacogenomics markers Genotyping of HLA-A and HLA-B was performed using HISAT-genotype31, xHLA32, and OptiType33, each applied independently to CRAM files with default parameters. Diplotypes were assigned by consensus: alleles concordant at two-field (four-digit) resolution in at least two tools were retained; discordant subtypes within the same allele group were collapsed to one-field (two-digit) resolution; and alleles without inter-method agreement were excluded. Only alleles meeting these criteria were considered in downstream analyses. We focused on the HLA-A and HLA-B loci, as these are the only HLA genes with pharmacogenetic markers currently supported by CPIC guidelines with actionable recommendations (evidence level A or B). To ensure a comprehensive assessment beyond CPIC, we additionally interrogated the HLA Adverse Drug Reaction (HLA-ADR) Database (http://www.allelefrequencies.net/; accessed 18 July 2025) to identify further drug–allele associations. From this database, we retained only markers reported with p < 0.01 across populations. Statistical analysis Allele frequency calculations Allele frequencies were calculated for each allele identified in the study. In cases of variants in the DPYD and RYR1 genes with uncertain haplotype configurations, linkage disequilibrium (LD) analysis was performed using the LDpair tool to evaluate potential patterns of co-inheritance34. Allele frequency comparisons across geogroups Allele frequency comparisons between the Moroccan and the reference populations (UK Biobank and Middle Eastern) were performed using chi-squared tests for common alleles (AF ≥ 0.05) and Fisher’s exact tests for rare alleles. Multiple testing correction was applied using the Bonferroni method, with statistical significance set at p = 0.05 (corrected). Clinically relevant alleles were cross-referenced with ClinPGx to identify variants with established dosing guidelines or phenotype associations. All plots and graphical representations were produced using R35 and Python36. Results Variants annotation The annotation of the remaining 26,828,731 genetic variants after filtrations revealed primarily single-nucleotide variants (n = 21,135,898), deletions (n = 2,672,315), and insertions (n = 2,561,983) (Fig. S1A). Most variants were located in intronic (n = 16,969,130) and intergenic (n = 6,111,655) regions (Fig. S1B). Among protein-coding variants, 2,684 were classified as high impact, 82,008 as moderate impact, and 71,095 as low impact (Fig. S1C). Functional prediction tools further annotated 21,749 variants as deleterious by SIFT and 11,626 as probably damaging by PolyPhen-2 (Fig. S1D). These findings highlight a substantial subset of potentially functional variants meriting further exploration, particularly in genes involved in drug metabolism and response. Pharmacogenomic variants in drug-related genes We identified 8,289 candidate genes with established roles in drug response, curated from DrugBank, ChEMBL, and ClinPGx databases. The functional classification of these genes revealed enzymes as the predominant category (4,189 genes, 50.53%), followed by membrane receptors (908 genes, 10.95%), transporters (327 genes, 3.94%), transcription factors (204 genes, 2.46%), ion channels (135 genes, 1.62%), other cytosolic proteins (115 genes, 1.38%), secreted proteins (102 genes, 1.23%), epigenetic regulators (40 genes, 0.48%), as well as many entries with unknown classification (2,269 genes ; 27.37%) (Fig. 1A). Within this gene set, we detected 4,991,339 variants, of which only 5,940 (0.1%) possessed documented pharmacogenomic associations in ClinPGx. A more stringent subset of 2,749 variants carried clinically relevant annotations pertaining to drug dosing or adverse reaction risk. The clinical evidence hierarchy revealed a pronounced bias toward lower-confidence classifications, with level 3 (n = 2,385 ; 87%) and level 4 (n = 157 ; 6%) annotations predominating, indicating limited or preliminary evidence. Conversely, high-confidence variants were remarkably scarce: level 1A (n = 19 ; 1%), level 1B (n = 7 ; 0.3%), level 2A (n = 17 ; 0.6%), and level 2B (n = 14 ; 0.5%). Gene-level analysis demonstrated that the majority of high-confidence variants were concentrated in a limited number of genes, with DPYD harboring the highest variant count (n = 9), followed by VKORC1 (n = 7). In contrast, several clinically important genes; such as CFTR, IFNL3, CYP2B6, CYP3A4, and ABCG2; contained fewer variants, ranging from one to three (Fig. 1B). These variants were mainly linked to toxicity (23 entries), efficacy (17 entries), dosage (12 entries), metabolism (5 entries), and other non-specific effects (3 entries) (Fig. 1C). Pharmacogenomic variation across the collected 8,289 drug-related genes. (A) Functional classification of drug-related genes based on ChEMBL annotations. (B) Distribution of variants with high-confidence clinical evidence (Levels 1–2) across the genes. (C) Categorization of evidence entries by phenotype type. (D) Distribution of novel variants identified in the clinically relevant pharmacogenes (CPIC), highlighting previously unreported pharmacogenomic variation in the Moroccan cohort. Of the total variants identified across the pharmacogenes, 129,021 (2.58%) were novel. The distribution of these variants was heterogeneous, with certain genes exhibiting a disproportionately high burden (Fig. 4B). DPYD harbored the highest number of novel variants (n = 187), followed by CYP2C19 (n = 30), SLCO1B1 (n = 27), CYP2C9 (n = 16), and CYP2C8 (n = 9) (Fig. 1D). Graphical representation of allele frequencies of Moroccan variants compared to reference populations. Statistical significance was assessed using chi-squared tests for common alleles (AF ≥ 0.05) and Fisher’s exact tests for rare alleles (AF < 0.05). Color intensity reflects p-value magnitude (darker = more significant). Values represent observed allele frequencies in each population. Potentially deleterious novel variants in drug-related genes Among the 129,021 novel variants identified across the selected pharmacogenes, 1,245 were predicted to be deleterious based on the CADD score (Phred > 20). Most of these variants had Phred scores between 20 and 30 (Fig. 2A), and 111 were classified as high impact (Fig. 2B). According to VEP functional annotations, the majority of these variants fell into the Missense & Protein Altering variants and Other categories, accounting for 1,108 variants in total (Fig. 2C; Supplementary Data 1). In addition, 99.8% of these variants had a minor allele frequency (AF) < 0.05, underscoring their rarity. Distribution of variants according to CADD Phred scores. (A) Histogram showing the total number of variants across Phred score bins. The red dashed line indicates the threshold of 20. (B) Boxplots comparing CADD Phred scores across VEP-defined impact categories. (C) Boxplots comparing CADD Phred scores across VEP-defined consequence categories. Native VEP consequences were further grouped into six classes (Protein Truncating Variants, Missense & Protein Altering, In-frame Indels, Synonymous, Splicing, and Other), based on the classification provided in Table S1. Circle symbols in boxplots represent outliers in each category. Chromosome 1 harbored the highest number of potentially deleterious novel variants (n = 123), whereas chromosome 21 contained the fewest (n = 7). These variants were enriched in relatively undercharacterized genes, with NPAS3 showing the greatest burden (n = 13). In contrast, genes with CPIC recommendations showed minimal enrichment for novel deleterious variants, with only RYR1 (n = 2) and DPYD (n = 1) affected (Supplementary Data 1). Clinically relevant pharmacogenomic variations in the Moroccan population The analysis of clinically relevant variants in the cohort revealed considerable allelic diversity and frequency variation (Fig. 3A), leading to genotypic and phenotypic heterogeneity that may sometimes warrant adjustments to clinical recommendations (Fig. 3B). Distribution of clinically relevant pharmacogenomic variations and predicted phenotypes in Moroccan individuals. (A) Frequencies of identified alleles across 17 genes in the cohort, with dot colors indicating allele function. (B) Graphical representation of pharmacogenomic phenotypes identified in the cohort and their associated therapeutic recommendations based on CPIC guidelines. The X-axis represents the prevalence of observed phenotypes in the cohort and the proportion of each related recommendation category, for each gene–drug combination shown on the Y-axis. The recommendations are classified into three categories: standard guide (no dosage adjustment required), alternate guide (requiring dosage modifications or alternative therapies), and no guide (no recommendation currently available in CPIC guidelines). Genes for which the star allele nomenclature is standardized are shown in red. No Function Alleles No Function Alleles The analysis revealed that the most common (AF ≥ 0.05) no function alleles in the Moroccan cohort were found in the CYP3A5, SLCO1B1, and CYP2C19 genes. In CYP3A5, the *3 allele was the most prevalent, with an allele frequency of 0.74, statistically comparable (p > 0.05) to those reported in Asian populations (Figs. 3A, 4; Table S2). For the same gene, the CYP3A5*6 allele also displayed a notable frequency (AF = 0.06), statistically similar to that observed in African American/Afro-Caribbean, Sub-Saharan African, and Middle Eastern populations (Fig. 3A; Fig. 4; Table S2). These two alleles were distributed across different genotypes, resulting in two main phenotypes, with a predominance of poor metabolizers (PM, 67.89%) over intermediate metabolizers (IM, 28.44%) (Supplementary data 2). According to CPIC guidelines for tacrolimus use, it is crucial to achieve an adequate drug concentration to ensure therapeutic efficacy while avoiding toxic levels that may lead to graft rejection. Accordingly, a higher initial tacrolimus dose is recommended for CYP3A5 expressers (EM and IM phenotypes). In contrast, nonexpressers (PMs) generally require a standard initial dose (Table 1). Table 1Clinically actionable genotypes and phenotypes commonly detected in the Moroccan cohort and their clinical implications.GeneGenotypes of the Moroccan individualsCorresponding phenotypesAffected drug/ drug class with published CPIC guidelines ( level A)Pgx on FDA labelCondition for the use of a drug or therapy areaCPIC Dosing recommendation (Moroccan population)CPIC Monitoring/AdjustmentCYP2D6*1/*41, *1/*1, *1/*2, *2/*2, *2 × 2/*41, *1/*5, *2/*2 × 2, *29/*41, *2 × 2/*68 + *4, *5/*17, *3/*41, *2/*43, *2/*17, *1/*43, *2/*68 + *4, *29/*29, *1/*2 × 2, *2/*2 × 3, *9/*28, *2/*41, *2/*2 × 4, *10/*68 + *4, *17/*35, *1/*4, *1/*31, *4/*29, *17/*41, *1/*1 × 2, *1/*68 + *4, *2/*45, *1 × 3/*2, *2 × 3/*41, *41/*41, *5/*45NM: 58,51 -NM(AS = 2): 44,68% -NM(AS = 1,25): 7,45%-NM(AS = 1,5): 5,32%-NM(AS = 2,25): 1,06%UM: 19,15% IM: 18,08-IM(AS = 1): 9,57%-IM(AS = 0,75): 3,19%-IM(AS = 0,5): 3,19%-IM(AS = 0,25): 2,13% Indeterminate: 4,25%Atomoxetine37Actionable PGxA non-stimulant therapy indicated for the treatment of attention-deficit/hyperactivity disorder (ADHD)NM, UM, IM (AS = 1): Start with 40 mg/day; increase to 80 mg/day after 3 days if no response and no adverse events after 2 weeks, consider 100 mg/dayIM (AS ≠ 1): Initiate with 40 mg/day. If no clinical response and no adverse events after 2 weeks, increase to 80 mg/dayNM, UM, IM (AS = 1): if no response after 2 weeks and no adverse events, measure peak plasma level (1–2 h post-dose); if < 200 ng/mL, increase dose proportionally to target 400 ng/mLIM (AS ≠ 1): if inadequate response after 2 weeks, obtain plasma concentration 2–4 h post-dose. If < 200 ng/mL, increase dose proportionally to approach 400 ng/mL. If unacceptable side effects occur at any time, consider a dose reductionOndansetron, Tropisetron38Ondansetron: Informative PGxChemotherapy-induced nausea and vomiting (CINV), postoperative nausea and vomiting (PONV)NM, IM: Initiate with standard recommended doseUM: Avoid Ondansetron and Tropisetron. Select an alternative 5-HT3 antagonist not primarily metabolized by CYP2D6 (e.g., Granisetron)Not specifiedTamoxifen39Actionable PGxHormone receptor–positive breast cancer (adjuvant therapy and prevention)NM, UM: Initiate with the standard dose (20 mg/day). Avoid moderate and strong CYP2D6 inhibitorsIM: Use an aromatase inhibitor as preferred over Tamoxifen regardless of CYP2D6 genotype. If contraindicated, consider Tamoxifen 40 mg/day (FDA approved)Not specifiedTCAs (Amitriptyline, Nortriptyline)40Actionable PGxDepression, neuropathic pain, fibromyalgia, migraine prophylaxis, sleep disorders (off-label), obsessive–compulsive disorder (OCD)NM: Initiate with the standard recommended doseUM: Avoid tricyclic use due to potential reduced efficacy; if necessary, titrate to a higher target dose than for normal metabolizersIM: reduction of the recommended starting dose by 25%UM, IM: Use therapeutic drug monitoring to guide dose adjustmentsSSRIs (Paroxetine), SSRI-like (Vortioxetine)41NADepression, anxiety disorders, and obsessive–compulsive disorder (OCD)UM:Paroxetine: Avoid useVortioxetine: Prefer an alternative not predominantly metabolized by CYP2D6; if used, start at the standard doseNM: Initiate with the standard recommended doseIM:Paroxetine: Consider a lower starting doseVortioxetine: Initiate with the standard recommended doseUM:Vortioxetine: Adjust dose according to efficacy and tolerabilityIM:Paroxetine: Consider a slower titration schedule compared with normal metabolizers to reduce the risk of adverse effectsOpiods(Codeine, Tramadol)42Codeine and Tramadol: Actionable PgxModerate to severe painUM: Avoid Codeine and Tramadol due to increased Morphine formation and toxicity risk. If opioid therapy is warranted, alternative agents should be consideredNM, IM: Use Codeine and Tramadol labels recommended age-specificor weight-specific dosingIM: If no response and opioid therapy is needed, alternative agents should be consideredCYP2C19*1/*1, *1/*17, *1/*2, *1/*35, *1/*38, *1/*9, *2/*10, *2/*17, *2/*2, *2/*24, *2/*35, *15/*17, *17/*17NM: 56,88%RM: 20,18%-RM(*1/*17): 19,27%IM: 15,60%PM: 2,75%UM: 2,75%LPM: 0,92%LIM: 0,92%Clopidogrel43Actionable PGxantiplatelet prodrug, commonly prescribed to reduce the risk of myocardial infarction and stroke in patients with acute coronarysyndromes and/or following percutaneous coronary interventionCardiovascular indications:NM, RM, UM: Use at standard recommended dose (75 mg/day)IM, LIM: avoid standard dose if possibleLPM, PM: Avoid clopidogrelNeurovascular indications:RM, UM: no recommendationsNM: Use at standard recommended dose (75 mg/day)IM, LIM: consider an alternative treatmentPM, LPM: Avoid ClopidogrelNot specifiedProton Pump Inhibitors (Omeprazole, Lansoprazole, Pantoprazole)44Gastroesophageal reflux disease (GERD), peptic ulcer disease, H. pylori eradication, NSAID prophylaxisUM: Increase starting daily dose by 100%NM, RM: Initiate standard dose. Consider increasing by 50–100% for H. pylori or erosive esophagitis treatmentIM, LIM, PM, LPM: Initiate standard dose. If therapy is chronic (> 12 weeks) and effective, consider 50% dose reductionUM, RM, NM: Monitor for efficacyIM, LIM, PM, LPM: Monitor for continued efficacy and potential toxicityVoriconazole45Invasive fungal infectionsUM, RM (*1/*17), PM: use an alternative treatmentNM , IM: Initiate with standard recommended doseNot specifiedTCAs (Amitriptyline)40NADepression, neuropathic pain, fibromyalgia, migraine prophylaxis, OCD, insomniaRM, UM, PM: Avoid tertiary amines due to increased metabolism and risk of subtherapeutic response. Prefer TCAs not metabolized by CYP2C19 (e.g., Nortriptyline, Desipramine)NM, IM: Initiate therapy with the recommended starting doseRMs, UMs: if a tertiary amine is used, apply therapeutic drug monitoring for dose adjustmentPMs: if a tertiary amine is used, start at 50% of the recommended dose and adjust with therapeutic drug monitoringSSRIs (Citalopram, Escitalopram, Sertraline)41Citalopram, Escitalopram: Actionable PgxDepression, anxiety disorders, OCDCitalopram, Escitalopram:RM, UM: Initiate with the standard starting dose. If inadequate response at maintenance dose, consider higher dose or switch to a non–CYP2C19-metabolized antidepressantNM: Initiate with the recommended starting doseIM, LIM: Initiate with recommended starting dose. Consider slower titration and possibly lower maintenance dosePM, LPM: consider an alternative treatment; if Citalopram or Escitalopram is used, start at a lower dose, titrate more slowly, and target 50% of the standard maintenance doseSertraline:RM, UM, NM: Initiate with standard starting doseIM, LIM: Initiate with the recommended starting dose. Consider slower titration and possibly lower maintenance dosePM, LPM: consider an alternative treatment; if Sertraline is used, start at a lower dose, titrate more slowly, and target 50% of the standard maintenance doseIM, RM, UM, LIM: Monitor for side effects, Drug-drug interactions, and account for patient-specific characteristicsLPM, PM: Monitor for QT prolongation; also limit to 20 mg/day in patients with hepatic impairment, those taking a CYP2C19 inhibitor, and patients > 60 years of ageFor Sertraline (all phenotypes), consider also the CYP2B6 metabolizer statusCYP2B6*1/*1, *1/*11, *1/*18, *1/*2, *1/*22, *1/*36, *1/*5, *1/*6, *1/*9, *2/*11, *2/*18, *2/*2, *2/*22, *2/*5, *2/*6, *2/*9, *4/*6, *4/*9, *5/*18, *5/*22, *5/*6, *6/*6IM: 44,95%NM: 41,28%PM: 6,42%Indeterminate: 7,34%Efavirenz46HIV-1 infectionNM: Initiate Efavirenz at standard dose (600 mg/day)IM: Consider reduced dose (400 mg/day) due to higher exposurePM: Consider reduced dose (400 or 200 mg/day) due to higher exposureIM: Monitor for CNS side effects (e.g., dizziness, vivid dreams)PM: Monitor for CNS side effects and potential treatmentdiscontinuationCYP2C9*1/*1, *1/*2, *1/*3, *1/*5, *1/*9, *1/*11, *2/*5, *9/*11NM: 83,49% IM: 16,51% -IM(AS = 1.5): 12,84% -IM (AS = 1): 3,67%NSAIDs (Celecoxib, Flurbiprofen, Lornoxicam,Iibuprofen, Meloxicam, Piroxicam, Tenoxicam)47Celecoxib, Flurbiprofen, Meloxicam, Piroxicam: Actionable PgxLornoxicam, Ibuprofen, Tenoxicam: NAPain, inflammation, feverNM and IM (AS = 1.5): Prescribe the standard NSAIDs doseIM (AS = 1): Use the lowest recommended starting dose and titrate cautiously. For Meloxicam, start at 50% of the lowest recommended dose and titrate cautiously up to 50% of the maximum recommended dose. Avoid the use of Piroxicam and TenoxicamMonitor for adverse eventsPhenytoin48Actionable PGxEpilepsy, seizure disordersHLA-B*15:02 negativeNM and IM (AS = 1.5): Use standard dosing; no initial dose adjustment neededIM (AS = 1): Use typical initial/loading dose; reduce subsequent maintenance doses by ~ 25%NM and IM (HLA-B*15:02 negative): Adjust according to therapeutic drug monitoring (TDM), clinical response, and side effects. Monitor for SJS/TEN as per standard careWarfarin49Actionable PgxAnticoagulation: prevention/treatment of thromboembolic disordersUse pharmacogenetic dosing algorithms to guide initial and maintenance Warfarin doses, incorporating the genotype of VKORC1 and CYP4F2 along with clinical factorsNot specifiedStatins (Fluvastatin)50NAHypercholesterolemia, prevention of cardiovascular eventsNM: Prescribe the desired starting dose and adjust based on disease-specific guidelinesIM: Prescribe ≤ 40 mg/day starting dose; if higher dose needed, consider alternative statin or combination therapyEvaluate potential drug-drug interactions, renal and hepatic function, before initiating statinIM: Carefully assess potential drug interactions and organ function, as pronounced interactions may elevate the risk of myopathyCYP3A5*1/*1, *1/*3, *1/*6, *3/*3, *3/*6, *3/*7, *6/*6, *6/*7PM: 67,89%IM: 28,44%EM:3,67%Tacrolimus51NAPrevention of organ transplant rejectionIM, EM: Increase starting dose by 1.5–2 × the standard dose (Max: 0.3 mg/kg/day)PM: Initiate with the standard starting doseUse therapeutic drug monitoring (TDM) to reach and maintain target Tacrolimus levelsCYP4F2*1/*1, *1/*2, *1/*3, *1/*4, *1/*5, *1/*6, *1/*13, *2/*4, *2/*5, *3/*3, *3/*4, *3/*5, *3/*6, *4/*5, *4/*6Indeterminate: 50,46%Normal dose phenotype (NDP): 25,69%Increased dose phenotype (IDP): 23,85%Warfarin52NAAnticoagulation: prevention/treatment of thromboembolic disordersUse pharmacogenetic dosing algorithms to guide initial and maintenance Warfarin doses, incorporating genotype of VKORC1 and CYP2C9 along with clinical factorsNot specifiedVKORC1C/C, C/T, T/T (C:reference allele, T/ variant allele with rs9923231)Decreased dose phenotype (DDP): 69,72%Normal dose phenotype (NDP): 30,28%Warfarin52Actionable PGxAnticoagulation: prevention/treatment of thromboembolic disordersUse pharmacogenetic dosing algorithms to guide initial and maintenance Warfarin doses, incorporating genotype of CYP2C9 and CYP4F2 along with clinical factorsNot specifiedIFNL3C/C, C/T, T/T (C: reference allele, T: rs2231142 variant)Favorable response (FavR): 44,95%Unfavorable response (UnfavR): 55,05%Peginterferon-alpha-based regimens53Peginterferon alfa-2a: NAPeginterferon alfa-2b: Actionable PgxTreat Chronic hepatitis C virus (HCV) infectionFavR: 70–90% SVR; supports use of PEG-IFN-α + RBV after 24–48 weeks of treatment; 80–90% of patients eligible for shortened therapy (24–28 weeks)UfavR:30–60% SVR; supports use of PEG-IFN-α + RBV after 24–48 weeks of treatment; 50% of patients eligible for shortened therapy (24–28 weeks)In both cases: consider implications before initiating PEG-IFN-α- and RBV-containingregimensNot specifiedDPYD*4/*9A + *5 + *6, *5/*5, *5/*6, *5/*9A + c.1129-5923C > G, *5/c.496A > G + *6, *6/*9A + c.496A > G, *9A/*5, *9A/*9A + c.1218G > A, *9A/*9A + c.496A > G, *9A/c.1218G > A, *9A/c.1371C > T, *9A/c.1896 T > C, *9A/c.525G > A + *6, *9A + *5 + *6/c.3067C > T, Reference/Reference, Reference/*4, Reference/*5, Reference/*6, Reference/*9A, Reference/*9A + *5, Reference/*9A + c.496A > G, Reference/*9A + c.496A > G + *6, Reference/c.1896 T > C, Reference/c.2846A > T, Reference/c.3067C > A, Reference/c.496A > GNM: 97,25%IM: 2,75%Fluoropyrimidines54Actionable PGxTreat solid tumors such as colorectal, breast, head and neck cancersNM: No dose adjustment needed; use label-recommended dosingIM: Reduce starting dose based on activity score, then titrate according to toxicity or therapeutic drug monitoring. (reduction by 25% to 50%)Not specifiedSLCO1B1*1/*1, *1/*14, *1/*15, *1/*20, *1/*27, *1/*31, *1/*37, *1/*41, *1/*46, *5/*14, *5/*37, *14/*15, *14/*20, *14/*27, *14/*37, *15/*15, *15/*20, *15/*37, *15/*41, *20/*20, *20/*37, *37/*37, *37/*41, *37/*46NF: 59,63%DF: 29,36%Indeterminate: 5,5%IF: 2,75%PF: 1,83%PDF: 0,92%Statins (Atorvastatin, Fluvastatin, Lovastatin, Pitavastatin, Pravastatin, Rosuvastatin, Simvastatin)50Rosuvastatin: Actionable PgxOther statins: NAHypercholesterolemia, cardiovascular preventionIF, NF: Prescribe desired starting dose; adjust based on disease-specific guidelinesDF, PDF:Atorvastatin, Pitavastatin:Initial dose recommendation ≤ 40 mg (Atorvastatin), ≤ 2 mg (Pitavastatin); adjust according to disease-specific guidelinesFluvastatin, Pravastatin:Prescribe desired starting dose; adjust according to disease-specific guidelines; increased risk of myopathy with doses > 40 mgRosuvastatin:Prescribe desired starting dose; adjust according to disease and population specific guidelines; increased risk of myopathy with doses > 20 mgLovastatin, Simvastatin:Prefer alternative statin; if used: ≤ 20 mg/day (Lovastatin), < 20 mg/day (Simvastatin)PF:Initial dose recommendation (adjust according to disease-specific guidelines):Atorvastatin, Rosuvastatin*: ≤ 20 mgPitavastatin: ≤ 1 mgFluvastatin, Pravastatin: ≤ 40 mgLovastatin, Simvastatin:Prefer alternative statin depending on the desired potencyEvaluate potential drug-drug interactions, renal and hepatic function, before initiating statinUGT1A1*1/*1, *1/*36, *1/*80 + *28, *36/*80 + *28, *37/*80, *80 + *28/*80 + *28IM: 50,46%NM: 41,28%PM: 7,34%Indeterminate: 0,92%Atazanavir55NAHIV infection (as part of antiretroviral therapy), potential for bilirubin elevationNM, IM: Atazanavir can be prescribed normally. Inform the patient that the jaundice risk is low (less than about a 1 in 20 chance of stopping Atazanavir because of jaundice)PM: An alternative medication should be considered due to the high risk of jaundice associated with AtazanavirClosely monitor for jaundiceRYR1Reference/Reference, [c.4024A > G + c.8360C > G]/Reference, c.4024A > G/Reference, c.6178G > T/Reference, [c.5360C > T + c.6178G > T]/Reference, c.11266C > G/Reference, c.14270G > A/Reference, c.6178G > T/c.6178G > T, c.8360C > G/Reference, c.4024A > G/c.6178G > T, c.10747G > C/ReferenceUncertain Susceptibility: 100%Volatile anesthetic agents (Evoflurane, Halothane, Enflurane, Isoflurane, Methoxyflurane, and Desflurane) and Succinylcholine56Desflurane, Enflurane, Isoflurane, Sevoflurane, Succinylcholine: Actionable PgxHalothane, Methoxyflurane: NAGeneral anesthesia, neuromuscular blockadeGenetic results do not exclude the risk of malignant hyperthermia (MH). About 50% of confirmed MH cases have no identifiable variantClinical judgment should guide anesthetic choice. Consider family history, clinical signs, or further testingNUDT15*1/*1NM: 100%Thiopurine57Testing RecommendedLeukemia, inflammatory bowel disease, and autoimmune conditionsNM: Follow the standard dosing recommendations for each thiopurineAllow at least 2 weeks to reach steady-state after each dose adjustmentCACNA1SReference/ReferenceUncertain Susceptibility: 100%Volatile anesthetic agents (Evoflurane, Halothane, Enflurane, Isoflurane, Methoxyflurane, and Desflurane) and Succinylcholine56Desflurane, Enflurane, Isoflurane, Sevoflurane, Succinylcholine: Actionable PgxHalothane, Methoxyflurane: NAGeneral anesthesia, neuromuscular blockadeThe genetic result does not exclude the risk of malignant hyperthermia (MH) . About 50% of confirmed MH cases have no identifiable variantClinical judgment should guide anesthetic choice. Consider family history, clinical signs, or further testingG6PDA/A, A/B (reference), A/MALE, A- 202A_376G/B (reference), B (reference)/B (reference), MALE/B (reference)Normal: 98,17%Variable: 1,83%Many drugs (Dapsone, Methylene blue, Nitrofurantoin(M), Pegloticase, Primaquine, Rasburicase, Tafenoquine, Toluidine blue)58Dapsone, Methylene blue, Nitrofurantoin(M): Actionable PgxPegloticase, Primaquine, Rasburicase, Tafenoquine: Testing RequiredToluidine blue: NAAcute hemolytic Anemia (AHA)Normal: Use standard dosing for drugs associated with G6PD sensitivityVariable:High-risk and medium-risk drugs, Primaquine:Ascertain G6PD status by measuring enzyme activity; guide drug use according to activity-based phenotype recommendationsLow-to-no risk drugs:No reason to avoid low-to-no risk drugs at standard doses regardless of G6PD statusRecommended monitoring for signs of chronic hemolysisABCG2G/G (G: rs2231142 reference)G/T (T: rs2231142 variant)NF: 95,41%DF: 4,59%Statins (Rosuvastatin)50NAHypercholesterolemia, cardiovascular preventionNF, DF: Prescribe standard starting dose; adjust according to disease- and population-specific guidelinesEvaluate potential drug-drug interactions, renal and hepatic function, before initiating statinTPMT*1/*1, *1/*3C, *1/*24, *8/*33NM: 98,66%Indeterminate: 4,59%IM: 2,75%Thiopurines (Mercaptopurine, Azathioprine, Thioguanine)57Testing RecommendedHypercholesterolemia, cardiovascular preventionNM: Follow the standard dosing recommendations for each thiopurineIM:Mercaptopurine, Azathioprine:Reduce starting dose by 30–80%Thioguanine:Reduce starting dose by 50–80%NM: Allow at least 2 weeks to reach steady-state after each dose adjustmentIM: Allow at least 2–4 weeks to reach steady-state after each dose adjustment Clinically actionable genotypes and phenotypes commonly detected in the Moroccan cohort and their clinical implications. NM, UM, IM (AS = 1): Start with 40 mg/day; increase to 80 mg/day after 3 days if no response and no adverse events after 2 weeks, consider 100 mg/day IM (AS ≠ 1): Initiate with 40 mg/day. If no clinical response and no adverse events after 2 weeks, increase to 80 mg/day NM, UM, IM (AS = 1): if no response after 2 weeks and no adverse events, measure peak plasma level (1–2 h post-dose); if < 200 ng/mL, increase dose proportionally to target 400 ng/mL IM (AS ≠ 1): if inadequate response after 2 weeks, obtain plasma concentration 2–4 h post-dose. If < 200 ng/mL, increase dose proportionally to approach 400 ng/mL. If unacceptable side effects occur at any time, consider a dose reduction NM, IM: Initiate with standard recommended dose UM: Avoid Ondansetron and Tropisetron. Select an alternative 5-HT3 antagonist not primarily metabolized by CYP2D6 (e.g., Granisetron) NM, UM: Initiate with the standard dose (20 mg/day). Avoid moderate and strong CYP2D6 inhibitors IM: Use an aromatase inhibitor as preferred over Tamoxifen regardless of CYP2D6 genotype. If contraindicated, consider Tamoxifen 40 mg/day (FDA approved) NM: Initiate with the standard recommended dose UM: Avoid tricyclic use due to potential reduced efficacy; if necessary, titrate to a higher target dose than for normal metabolizers IM: reduction of the recommended starting dose by 25% UM: Paroxetine: Avoid use Vortioxetine: Prefer an alternative not predominantly metabolized by CYP2D6; if used, start at the standard dose NM: Initiate with the standard recommended dose IM: Paroxetine: Consider a lower starting dose Vortioxetine: Initiate with the standard recommended dose UM: Vortioxetine: Adjust dose according to efficacy and tolerability IM: Paroxetine: Consider a slower titration schedule compared with normal metabolizers to reduce the risk of adverse effects UM: Avoid Codeine and Tramadol due to increased Morphine formation and toxicity risk. If opioid therapy is warranted, alternative agents should be considered NM, IM: Use Codeine and Tramadol labels recommended age-specific or weight-specific dosing NM: 56,88% RM: 20,18% -RM(*1/*17): 19,27% IM: 15,60% PM: 2,75% UM: 2,75% LPM: 0,92% LIM: 0,92% antiplatelet prodrug, commonly prescribed to reduce the risk of myocardial infarction and stroke in patients with acute coronary syndromes and/or following percutaneous coronary intervention Cardiovascular indications: NM, RM, UM: Use at standard recommended dose (75 mg/day) IM, LIM: avoid standard dose if possible LPM, PM: Avoid clopidogrel Neurovascular indications: RM, UM: no recommendations NM: Use at standard recommended dose (75 mg/day) IM, LIM: consider an alternative treatment PM, LPM: Avoid Clopidogrel UM: Increase starting daily dose by 100% NM, RM: Initiate standard dose. Consider increasing by 50–100% for H. pylori or erosive esophagitis treatment IM, LIM, PM, LPM: Initiate standard dose. If therapy is chronic (> 12 weeks) and effective, consider 50% dose reduction UM, RM, NM: Monitor for efficacy IM, LIM, PM, LPM: Monitor for continued efficacy and potential toxicity UM, RM (*1/*17), PM: use an alternative treatment NM , IM: Initiate with standard recommended dose RM, UM, PM: Avoid tertiary amines due to increased metabolism and risk of subtherapeutic response. Prefer TCAs not metabolized by CYP2C19 (e.g., Nortriptyline, Desipramine) NM, IM: Initiate therapy with the recommended starting dose RMs, UMs: if a tertiary amine is used, apply therapeutic drug monitoring for dose adjustment PMs: if a tertiary amine is used, start at 50% of the recommended dose and adjust with therapeutic drug monitoring Citalopram, Escitalopram: RM, UM: Initiate with the standard starting dose. If inadequate response at maintenance dose, consider higher dose or switch to a non–CYP2C19-metabolized antidepressant NM: Initiate with the recommended starting dose IM, LIM: Initiate with recommended starting dose. Consider slower titration and possibly lower maintenance dose PM, LPM: consider an alternative treatment; if Citalopram or Escitalopram is used, start at a lower dose, titrate more slowly, and target 50% of the standard maintenance dose Sertraline: RM, UM, NM: Initiate with standard starting dose IM, LIM: Initiate with the recommended starting dose. Consider slower titration and possibly lower maintenance dose PM, LPM: consider an alternative treatment; if Sertraline is used, start at a lower dose, titrate more slowly, and target 50% of the standard maintenance dose IM, RM, UM, LIM: Monitor for side effects, Drug-drug interactions, and account for patient-specific characteristics LPM, PM: Monitor for QT prolongation; also limit to 20 mg/day in patients with hepatic impairment, those taking a CYP2C19 inhibitor, and patients > 60 years of age For Sertraline (all phenotypes), consider also the CYP2B6 metabolizer status IM: 44,95% NM: 41,28% PM: 6,42% Indeterminate: 7,34% NM: Initiate Efavirenz at standard dose (600 mg/day) IM: Consider reduced dose (400 mg/day) due to higher exposure PM: Consider reduced dose (400 or 200 mg/day) due to higher exposure IM: Monitor for CNS side effects (e.g., dizziness, vivid dreams) PM: Monitor for CNS side effects and potential treatment discontinuation Celecoxib, Flurbiprofen, Meloxicam, Piroxicam: Actionable Pgx Lornoxicam, Ibuprofen, Tenoxicam: NA NM and IM (AS = 1.5): Prescribe the standard NSAIDs dose IM (AS = 1): Use the lowest recommended starting dose and titrate cautiously. For Meloxicam, start at 50% of the lowest recommended dose and titrate cautiously up to 50% of the maximum recommended dose. Avoid the use of Piroxicam and Tenoxicam HLA-B*15:02 negative NM and IM (AS = 1.5): Use standard dosing; no initial dose adjustment needed IM (AS = 1): Use typical initial/loading dose; reduce subsequent maintenance doses by ~ 25% NM: Prescribe the desired starting dose and adjust based on disease-specific guidelines IM: Prescribe ≤ 40 mg/day starting dose; if higher dose needed, consider alternative statin or combination therapy Evaluate potential drug-drug interactions, renal and hepatic function, before initiating statin IM: Carefully assess potential drug interactions and organ function, as pronounced interactions may elevate the risk of myopathy PM: 67,89% IM: 28,44% EM:3,67% IM, EM: Increase starting dose by 1.5–2 × the standard dose (Max: 0.3 mg/kg/day) PM: Initiate with the standard starting dose Indeterminate: 50,46% Normal dose phenotype (NDP): 25,69% Increased dose phenotype (IDP): 23,85% Decreased dose phenotype (DDP): 69,72% Normal dose phenotype (NDP): 30,28% Favorable response (FavR): 44,95% Unfavorable response (UnfavR): 55,05% Peginterferon alfa-2a: NA Peginterferon alfa-2b: Actionable Pgx FavR: 70–90% SVR; supports use of PEG-IFN-α + RBV after 24–48 weeks of treatment; 80–90% of patients eligible for shortened therapy (24–28 weeks) UfavR:30–60% SVR; supports use of PEG-IFN-α + RBV after 24–48 weeks of treatment; 50% of patients eligible for shortened therapy (24–28 weeks) In both cases: consider implications before initiating PEG-IFN-α- and RBV-containing regimens NM: 97,25% IM: 2,75% NM: No dose adjustment needed; use label-recommended dosing IM: Reduce starting dose based on activity score, then titrate according to toxicity or therapeutic drug monitoring. (reduction by 25% to 50%) NF: 59,63% DF: 29,36% Indeterminate: 5,5% IF: 2,75% PF: 1,83% PDF: 0,92% Rosuvastatin: Actionable Pgx Other statins: NA IF, NF: Prescribe desired starting dose; adjust based on disease-specific guidelines DF, PDF: Atorvastatin, Pitavastatin: Initial dose recommendation ≤ 40 mg (Atorvastatin), ≤ 2 mg (Pitavastatin); adjust according to disease-specific guidelines Fluvastatin, Pravastatin: Prescribe desired starting dose; adjust according to disease-specific guidelines; increased risk of myopathy with doses > 40 mg Rosuvastatin: Prescribe desired starting dose; adjust according to disease and population specific guidelines; increased risk of myopathy with doses > 20 mg Lovastatin, Simvastatin: Prefer alternative statin; if used: ≤ 20 mg/day (Lovastatin), < 20 mg/day (Simvastatin) PF: Initial dose recommendation (adjust according to disease-specific guidelines): Atorvastatin, Rosuvastatin*: ≤ 20 mg Pitavastatin: ≤ 1 mg Fluvastatin, Pravastatin: ≤ 40 mg Lovastatin, Simvastatin: Prefer alternative statin depending on the desired potency IM: 50,46% NM: 41,28% PM: 7,34% Indeterminate: 0,92% NM, IM: Atazanavir can be prescribed normally. Inform the patient that the jaundice risk is low (less than about a 1 in 20 chance of stopping Atazanavir because of jaundice) PM: An alternative medication should be considered due to the high risk of jaundice associated with Atazanavir Desflurane, Enflurane, Isoflurane, Sevoflurane, Succinylcholine: Actionable Pgx Halothane, Methoxyflurane: NA Desflurane, Enflurane, Isoflurane, Sevoflurane, Succinylcholine: Actionable Pgx Halothane, Methoxyflurane: NA Normal: 98,17% Variable: 1,83% Dapsone, Methylene blue, Nitrofurantoin(M): Actionable Pgx Pegloticase, Primaquine, Rasburicase, Tafenoquine: Testing Required Toluidine blue: NA Normal: Use standard dosing for drugs associated with G6PD sensitivity Variable: High-risk and medium-risk drugs, Primaquine: Ascertain G6PD status by measuring enzyme activity; guide drug use according to activity-based phenotype recommendations Low-to-no risk drugs: No reason to avoid low-to-no risk drugs at standard doses regardless of G6PD status G/G (G: rs2231142 reference) G/T (T: rs2231142 variant) NF: 95,41% DF: 4,59% NM: 98,66% Indeterminate: 4,59% IM: 2,75% NM: Follow the standard dosing recommendations for each thiopurine IM: Mercaptopurine, Azathioprine: Reduce starting dose by 30–80% Thioguanine: Reduce starting dose by 50–80% NM: Allow at least 2 weeks to reach steady-state after each dose adjustment IM: Allow at least 2–4 weeks to reach steady-state after each dose adjustment On the other hand, the SLCO1B1*15 was the most common SLCO1B1 allele with a frequency of 0.14 (Figs. 3A, 4; Table S2). This value is statistically comparable to those in the East Asian, European, and Middle Eastern populations (Fig. 4; Table S2). Among carriers of this allele, 23.85% exhibited a decreased function (DF), 1.83% a poor function (PF), and 0.92% a possible decreased function (PDF) phenotype. The CPIC recommendations for these phenotypes involve adjusting the statin dose (based on disease and population specific guidelines) or prescribing alternative statins (Table 1). Finally, for the CYP2C19 gene, the *2 allele was the most frequent non-functional allele (AF = 0.09) identified in the Moroccan population (Figs. 3A, 4; Table S2). Its frequency is statistically similar to those identified in the African American/Afro-Caribbean, European, Sub-Saharan African, and Middle Eastern populations (Fig. 4; Table S2). Individuals carrying this allele exhibited the following phenotypes: intermediate metabolizers (IMs, 13.76%), poor metabolizers (PMs, 2.75%), and likely poor metabolizers (LPMs, 0.92%) (Supplementary data 2). Thus, for IMs, PMs, and LPMs, the CPIC guidelines recommend dose reduction or alternative therapies, particularly for clopidogrel, selective serotonin reuptake inhibitors (SSRIs), and proton pump inhibitors (PPIs). In the case of voriconazole and the tricyclic antidepressant amitriptyline, these recommendations apply specifically to individuals with the PM phenotype (Table 1). Decreased Function Alleles Decreased Function Alleles Decreased function alleles with AF ≥ 0.05 were detected across multiple genes, notably UGT1A1, CYP4F2, CYP2B6, CYP2D6, and CYP2C9. In our cohort, the UGT1A1*80 + *28 allele was observed at an allele frequency of 0.33, statistically comparable to frequencies reported in other populations, except for the East Asian population, where it is significantly lower (Figs. 3A, 4; Table S2). Individuals carrying this allele were predominantly classified as intermediate metabolizers (50.46%) and, to a lesser extent, poor metabolizers (7.34%) (Supplementary data 2). According to CPIC guidelines for atazanavir, a widely prescribed antiretroviral, standard dosing is appropriate for IMs; for PMs, alternative therapies should be considered (Table 1). Another notable allele observed in our cohort was the CYP4F2*3 allele, with a frequency of 0.22 (Figs. 3A, 4; Table S2). Significant differences were noted compared to the African American/Afro-Caribbean and Sub-Saharan African populations, where the allele frequency was markedly lower (Fig. 4; Table S2). Among carriers of this allele, 23.85% exhibited an increased dose phenotype, while 15.6% had an indeterminate phenotype (Supplementary data 2). According to current pharmacogenetic guidelines, incorporating CYP4F2*3 alongside CYP2C9, VKORC1, and relevant clinical parameters into warfarin dosing algorithms is especially recommended for individuals of non-African ancestry. In such cases, the presence of the CYP4F2*3 allele typically justifies a higher initial warfarin dose to achieve optimal anticoagulation (Table 1). Regarding the CYP2B6 gene, the most frequent decreased function alleles identified were *6 and *9, with allele frequencies of 0.21 and 0.06, respectively (Fig. 3A). Carriers of these alleles were classified as intermediate metabolizers (39.45%) and poor metabolizers (6.42%) (Supplementary data 2). For these individuals, clinical guidelines recommend initiating efavirenz therapy (an antiretroviral agent commonly prescribed as part of the treatment of HIV infection) at reduced doses to minimize the risk of central nervous system toxicity (Table 1). The CYP2D6 gene plays a key role in the metabolism of a broad range of medications. In our cohort, the most prevalent CYP2D6 decreased function allele was *41, with an allele frequency of 0.08 (Figs. 3A, 4; Table S2). This frequency is statistically comparable to those reported in Asian populations, notably Central/South Asians and East Asians (Fig. 4; Table S2). Among CYP2D6*41 carriers, the observed phenotypes included intermediate metabolizers (IMs; 5.32%), normal metabolizers (NMs; 8.51%), and ultra-rapid metabolizers (UMs; 1.06%) (Supplementary data 2). The clinical relevance of CYP2D6 polymorphisms is underscored by several CPIC guidelines. For instance, in individuals classified as IMs (Activity Score (AS) ≠ 1), atomoxetine metabolism is reduced, leading to elevated plasma concentrations. It is recommended in such cases to monitor patients closely for adverse effects and to adjust the dose downward if necessary (Table 1). Moreover, CYP2D6 variations can influence the efficacy of antiemetics such as ondansetron and tropisetron, which are commonly prescribed to prevent nausea and vomiting associated with chemotherapy, radiotherapy, or surgery. In ultra-rapid metabolizers, these agents may be less effective due to accelerated metabolism; therefore, alternative medications not primarily metabolized by CYP2D6 should be considered (Table 1). Opioid analgesics such as codeine and tramadol are also substrates of CYP2D6. For UM individuals, their use is not recommended due to the increased risk of toxicity resulting from their rapid conversion into active metabolites. If opioid therapy is warranted, alternative agents that are not primarily metabolized by CYP2D6 should be considered. For intermediate metabolizers, these opioids may be used with caution. However, if there is insufficient analgesic response and opioid treatment is warranted, switching to non-CYP2D6-dependent alternatives is advised (Table 1). In addition to analgesics, the CYP2D6 genotype also impacts the metabolism of certain antidepressants. For instance, the use of some tricyclic antidepressants (TCAs) and paroxetine is not recommended for UMs due to the risk of therapeutic failure. Otherwise, if a TCA is warranted, a higher target dose may be required, accompanied by therapeutic drug monitoring. For IMs, reducing the initial dose should be considered to minimize the risk of adverse effects. On the other hand, for vortioxetine, an alternative antidepressant not predominantly metabolized by CYP2D6, is preferable in UMs. However, if its use is clinically justified, treatment can be initiated at the standard dose, with subsequent adjustment of the maintenance dose based on therapeutic response (Table 1). Finally, for IM individuals receiving tamoxifen, hormonal therapy such as an aromatase inhibitor is generally recommended as a preferred alternative. If this option is not indicated, a higher tamoxifen dose (40 mg/day) may be considered, with caution to avoid concomitant administration of CYP2D6 strong to weak inhibitors (Table 1). Another key gene is CYP2C9, which metabolizes several drugs. The most common decreased function allele identified was CYP2C9*2, with an allele frequency of 0.05, statistically comparable to that observed in the African American/Afro-Caribbean, Central/South Asian, European, and Middle Eastern populations (Figs. 3A, 4; Table S2). All individuals carrying this allele were classified as IMs (10.09%) (Supplementary data 2). For these individuals, the CPIC provides specific therapeutic recommendations depending on the drug involved. For warfarin, a dose reduction combined with the use of dosing algorithms is recommended due to the risk of bleeding during warfarin therapy. For phenytoin, a reduced maintenance dose is suggested for IMs with an AS of 1.0, with subsequent adjustments guided by therapeutic drug monitoring. Regarding fluvastatin, the doses above 40 mg should be avoided, and an alternative statin may be considered to achieve the desired efficacy. Lastly, for NSAIDs, particularly in IMs with an AS of 1.0, the recommendation is to start with the lowest effective dose for the shortest duration, with careful monitoring for adverse events. For meloxicam, treatment should be initiated at 50% of the lowest recommended dose and titrated cautiously. The use of piroxicam and tenoxicam is not recommended (Table 1). Increased Function Alleles Increased Function Alleles The analysis also examined increased function alleles associated with higher enzymatic activity. The most frequently observed variants were found in the CYP2C19, SLCO1B1, and CYP2D6 genes. Among these, CYP2C19*17 was the most common allele, with an allele frequency of 0.15, significantly lower than that of the East Asian population (AF = 0.008) (Figs. 3A, 4; Table S2). Carriers of this allele were classified as rapid (20.18%), intermediate (3.67%), and ultra-rapid metabolizers (2.75%) (Supplementary data 2). According to CPIC guidelines, either an increase in dosage or the use of alternative therapies is recommended for rapid and ultra-rapid metabolizers, particularly with certain SSRIs such as citalopram and escitalopram, PPIs, voriconazole, and TCAs (Table 1). The most frequent increased function alleles in the SLCO1B1 gene are *14 and *20. Allele *14 has an AF of 0.10, statistically comparable to that reported in European (AF = 0.07) and Middle Eastern (AF = 0.13) populations. In addition, allele *20 has a frequency of 0.09, significantly higher than that observed in Asian populations (CAS/SAS: AF = 0.02; EAS: AF = 0.002) (Figs. 3A, 4; Table S2). Carriers of these alleles were classified into different functional phenotypes: normal function (21.10%), decreased function (5.50%), increased function (2.75%), and indeterminate function (0.92%) (Supplementary data 2). Standard recommendations must be followed for individuals with normal or increased function phenotypes. However, for those with a decreased function phenotype, CPIC guidelines suggest dose adjustments or prescribing alternative statins (Table 1). The last prevalent increased function allele is *2 × 2 (a duplication of the *2 allele) in the CYP2D6 gene, with an allele frequency of 0.09. Individuals with this allele exhibited an ultra-rapid metabolizer phenotype (13.83%) or a normal metabolizer phenotype (3.19%) (Supplementary data 2). The clinical relevance of these phenotypes has been reported in the previous section. Other prevalent alleles with functional relevance Other prevalent alleles with functional relevance We identified additional clinically relevant variants in other pharmacogenes within the Moroccan population, notably the T allele (rs9923231) of the VKORC1 gene which was observed at a high frequency in the cohort (AF = 0.47). This frequency is comparable to that reported in European (AF = 0.38) and Middle Eastern (AF = 0.50) populations (Figs. 3A, 4; Table S2). Among individuals carrying this allele (69.72%), all exhibited a phenotype associated with increased sensitivity to warfarin, requiring reduced dosing (Supplementary data 2). Indeed, patients who are heterozygous or homozygous for the T allele require significantly lower doses than C/C homozygotes. This genetic variant should be considered in warfarin dosing algorithms, in association with CYP2C9 and CYP4F2, to improve the safety and efficacy of the therapy (Table 1). In addition, another variant of interest was identified: the T allele (rs12979860) of the IFNL3 gene, with an allele frequency of 0.33, statistically similar to frequencies observed in European (AF = 0.30) and Central/South Asian (AF = 0.24) populations (Figs. 3A, 4; Table S2). This variant is associated with a reduced response to hepatitis C treatment with pegylated interferon-α (PEG-IFN-α) and ribavirin (RBV) in individuals carrying at least one copy of the T allele (55.05%). Compared to individuals with a favorable response genotype (44.95%), these patients have a reduced chance of achieving sustained virological response (SVR) (Supplementary data 2) (Table 1). As such, this pharmacogenetic information can inform shared decision-making between clinicians and patients, particularly considering the significant side effects associated with this therapy. Overall, most pharmacogenes exhibited a predominance of normal function alleles. For instance, NUDT15 and CACNA1S exclusively carried normal function alleles in the entire cohort. Similarly, a clear predominance of normal function alleles was observed for TPMT, RYR1, CYP2C9, UGT1A1, ABCG2, CYP2C19, and G6PD, with allele frequencies exceeding 0.5 (Figs. 3A, 4; Table S2). HLA-A and HLA-B allelic and genotypic diversity HLA typing of Moroccan individuals revealed 32 and 43 distinct HLA-A and HLA-B alleles respectively, showing a marked allelic diversity (Supplementary data 3). Among the HLA-A alleles, five exhibited AFs ≥ 0.05: A*03:01 (0.2), A*01:01 (0.11), A*30:02 (0.11), A*30:01 (0.08), and A*03:02 (0.08). These alleles contributed to 15 distinct genotypes, the most frequent being A*03:01/A*03:01 (18.3%), which collectively accounted for 62.42% of the cohort. For HLA-B, the most frequent alleles were B*08:01 (0.1), B*07:02 (0.1), B*51:01 (0.09), B*14:02 (0.07), and B*38:01 (0.05) (Supplementary data 3). These alleles were involved in 18 distinct genotypes, with B*07:02/B*07:02 being the most frequent (8.05%), together representing 48.3% of the cohort. Prevalence of HLA-A and HLA-B pharmacogenomic markers The pharmacogenomic analysis identified 19 clinically relevant HLA alleles, comprising 10 HLA-A and 9 HLA-B variants, significantly associated with adverse drug reactions. Cross-referencing with the HLA-ADR database revealed 64 potential drug-HLA-ADR associations encompassing 19 distinct drugs and 12 different ADRs. Among these, HLA-A*11:01 (AF = 0.02) showed the highest number of associations (n = 9), predominantly linked to carbamazepine-induced severe cutaneous adverse reactions (SCARs) including acute generalized exanthematous pustulosis (AGEP), drug-induced hypersensitivity syndrome (DIHS), Stevens-Johnson syndrome (SJS), toxic epidermal necrolysis (TEN,) and severe ocular complications. The HLA-A*33:01 (AF = 0.02) allele was exclusively associated with drug-induced liver injury (DILI) across 7 medications including enalapril, erythromycin, fenofibrate, methyldopa, sertraline, terbinafine, and ticlopidine. On the other hand, HLA-B*51:01 (AF = 0.09) demonstrated 6 associations, particularly with antiepileptic drugs, clindamycin, and phenytoin, leading to AGEP, DIHS, drug-induced systemic illness (DISI), SJS, and TEN. HLA-A*31:01 (AF = 0.02) and HLA-A*02:01 (AF = 0.03) each showed 5 associations, with A*31:01 linked to antiepileptic drug-induced AGEP, DIHS, maculopapular exanthema (MPE), SJS, and TEN, and A*02:01 associated with amoxicillin-clavulanate and antibacterial drug reactions causing AGEP, DIHS, DILI, SJS, and TEN. HLA-A*33:03 (AF = 0.01) was specifically associated with allopurinol, carbamazepine, and ticlopidine-induced cholestatic DILI, SJS, and TEN. Despite their relatively high frequencies, some alleles showed limited associations. For instance, HLA-A*30:02 (AF = 0.11) was specifically linked to amoxicillin–clavulanate–induced DILI with hepatocellular-type damage, while HLA-A*03:01 (AF = 0.20) was associated exclusively with phenytoin-induced fixed drug eruption (FDE). The remaining alleles (HLA-B*38:01, B*13:02, B*44:03, A*66:01, B*57, B*58:01, A*80:01, B*40:01, A*32:01, B*07:05, and B*40:02) demonstrated 1–4 associations each, with frequencies ranging from 0.01 to 0.05, collectively encompassing a broad spectrum of medications including lamotrigine, allopurinol, NSAIDs, vancomycin, nevirapine, and oxcarbazepine, and associated adverse reactions spanning SJS, TEN, DIHS, DISI, MPE, and cholestatic DILI (Table S3; Fig. 5). Population-specific immunogenetic risk for ADRs. (A) Heatmap of significant associations between HLA alleles and specific ADR phenotypes, revealing phenotype-specific susceptibility patterns. (B) Drug-level immunogenetic risk burden, quantified as the number of HLA allele associations per drug/drug class, highlighting therapeutics with the broadest immunogenic potential. Discussion In this study, the genomic analysis of a Moroccan cohort provides insights into the population-specific pharmacogenomic variation and challenges conventional clinical practice. Our findings showed a limited number of high-confidence pharmacogenomic annotations in our data (Levels 1 and 2, representing only 2.4% of clinically relevant variants), further exposing the challenges of translating genomic discoveries into clinical applications. This bias toward low-confidence classifications reflects the historical underrepresentation of diverse populations in pharmacogenomic research. Consequently, we face a paradox: populations such as North Africans, with high genetic diversity, have the least clinical guidance59. Thus, we identified 129,021 novel variants, many of which are concentrated in clinically important genes such as DPYD, CYP2C19, and SLCO1B1. Furthermore, the analysis of these novel variants using CADD identified 1,245 potentially deleterious variants. Genes with CPIC recommendations were unaffected, except for three variants in RYR1 and DPYD . The presence of potentially damaging novel variants in genes such as DPYD is particularly concerning, given its critical role in modulating 5-fluorouracil toxicity and the risk of severe adverse effects in cancer patients. Notably, several genes that remain poorly investigated in the context of pharmacogenomics harbored potentially deleterious novel variants, including NPAS3 (13 variants), which has previously been associated with the response to certain antipsychotic drugs60. Therefore, such genes warrant deeper investigation in the Moroccan population. The analysis of functionally significant variants (according to CPIC) in our cohort reveals that some alleles are widespread. For instance, the SLCO1B1*15 haplotype, which includes the c.521 T > C variant (rs4149056, p.Val174Ala), is notable for its reduction of the transporter function, resulting in decreased clearance of several drugs such as simvastatin, and a higher risk of myopathy in carriers61. Similarly, UGT1A1*80 + *28 which include the polymorphism UGT1A1*28, characterized by seven thymine-adenine repeats in the promoter region of the UGT1A1 gene, initially identified as the causal genetic variant of Gilbert syndrome, a benign form of unconjugated hyperbilirubinemia55,62. On the other hand, CYP2C19*17 is associated with gene overexpression and increased enzymatic activity. Carriers of this allele may exhibit an ultra-rapid metabolizer phenotype, often requiring higher drug doses to achieve the intended therapeutic effect63,64. Finally, the T allele (rs9923231) of the VKORC1 gene, also highly prevalent in our cohort, is linked to lower warfarin dose requirements and increased drug sensitivity65. These findings underscore the need to establish a minimum panel of common pharmacogenomic variants tailored to the Moroccan population to guide safer and more effective prescribing practices. Although rare variants can substantially contribute to pharmacogenomic variability and influence pharmacogenomic recommendations, their integration into routine clinical practice remains challenging, as it requires comprehensive functional characterization prior to incorporation into preemptive pharmacogenetic testing strategies66. ADRs represent a leading cause of hospitalization and mortality in healthcare worldwide. SCARs such as SJS and TEN have mortality rates ranging from 10 to 30%67. Our HLA genotyping highlights population-specific allelic diversity with important implications for ADR risk assessment. The high frequencies of alleles such as HLA-A*03:01, HLA-A*30:02, and HLA-B*51:01 in this Moroccan cohort may confer distinct susceptibility profiles to drug-induced hypersensitivity reactions. SCARs, particularly SJS and TEN, were predominant in our population, requiring appropriate pharmacogenomic screening. Moreover, the 64 potential drug–HLA associations identified, spanning 19 drugs and drug classes and 12 adverse reaction phenotypes, represent a substantial clinical burden that remains unrecognized in current prescribing practices. Despite these significant insights, certain limitations must be acknowledged. The relatively modest sample size may not fully capture Morocco’s complex demographic and ethnic diversity. Furthermore, the absence of individual phenotypic drug response data prevents direct validation of genotype–phenotype correlations. In addition, experimental long-read sequencing is planned for Phase 2 of the Moroccan Genome Project, enabling improved detection of structural variants and more comprehensive genomic characterization. Nonetheless, this first phase of the Moroccan Genome Project, representing approximately 10% of the targeted 1000 genomes, provides a critical foundation for advancing pharmacogenomics in Morocco and North Africa12. Conclusion These results call for a paradigm shift toward inclusive pharmacogenomics that reflects global genetic diversity. Precision medicine can only achieve its full potential if underrepresented populations are systematically integrated into research frameworks. The pharmacogenomic profile of this Moroccan cohort highlights the shortcomings of Eurocentric approaches and provides a critical foundation for implementing population-specific pharmacogenomics in Morocco and across North Africa. Expanding genomic diversity in pharmacogenomic research is no longer optional but necessary to ensure that the benefits of precision medicine extend to all populations, fulfilling its fundamental promise of optimized, individualized therapy. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Acknowledgements The authors gratefully acknowledge the Mohammed VI Center for Research and Innovation (CM6RI), under the Mohammed VI Foundation for Sciences and Health, for its institutional and logistical support that facilitated the completion of this research. We also thank the Clinical Pharmacogenetics Implementation Consortium (CPIC) for providing publicly available pharmacogenomic data and resources that supported this work. Author contributions SK: Conceptualization, Methodology, Data curation, Formal Analysis, Investigation, Visualization, Writing—original draft. MWC: Methodology, Data curation, Formal Analysis, Investigation, Visualization, Writing—original draft. MH: Methodology, Data curation, Formal Analysis, Investigation, Methodology, Visualization, Writing—original draft. MA: Investigation, Data curation, review and editing. REJ: review and editing. LBA: review and editing. EEF: review and editing. LBE: Supervision, review and editing, Funding acquisition. SB: Project administration, Methodology, Supervision, review and editing. Funding This research did not receive funding. Data availability Raw sequencing data and the VCF file are available via EGA under the study ID (EGAC50000000353). Researchers must first apply through the EGA data access committee (DAC), following the standard EGA procedure outlined on their website. Declarations Competing interests The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Mohammed Walid Chemao-Elfihri and Mohammed Hakmi have equally contributed to this work. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Data Availability Statement Raw sequencing data and the VCF file are available via EGA under the study ID (EGAC50000000353). Researchers must first apply through the EGA data access committee (DAC), following the standard EGA procedure outlined on their website.