Original Article | Published:

Association of common genetic variants of HOMER1 gene with levodopa adverse effects in Parkinson’s disease patients

The Pharmacogenomics Journal volume 14, pages 289294 (2014) | Download Citation


Levodopa is the most effective symptomatic therapy for Parkinson’s disease, but its chronic use could lead to chronic adverse outcomes, such as motor fluctuations, dyskinesia and visual hallucinations. HOMER1 is a protein with pivotal function in glutamate transmission, which has been related to the pathogenesis of these complications. This study investigates whether polymorphisms in the HOMER1 gene promoter region are associated with the occurrence of the chronic complications of levodopa therapy. A total of 205 patients with idiopathic Parkinson’s disease were investigated. Patients were genotyped for rs4704559, rs10942891 and rs4704560 by allelic discrimination with Taqman assays. The rs4704559 G allele was associated with a lower prevalence of dyskinesia (prevalence ratio (PR)=0.615, 95% confidence interval (CI) 0.426–0.887, P=0.009) and visual hallucinations (PR=0.515, 95% CI 0.295–0.899, P=0.020). Our data suggest that HOMER1 rs4704559 G allele has a protective role for the development of levodopa adverse effects.


Idiopathic Parkinson’s disease (PD) affects 1–3% of people older than 65 years, and its core symptoms are slowness of movements, rigidity and resting tremor.1 The complete pathological pathway underlying this condition remains unknown, but the mechanism involved in the cardinal motor manifestations of disease results from a progressive loss of dopaminergic cells in the midbrain, leading to a dopamine deficit in the striatum. Levodopa is a dopamine precursor, and it is used as the major pharmacological agent to restore this deficiency. Despite the important symptomatic effect provided by levodopa, its long-term use is associated with adverse outcomes such as motor fluctuations, dyskinesia and visual hallucinations, which could severely impair daily life.2, 3

Motor fluctuation is a phenomenon characterized by a predictable (wearing-off) or an unpredictable (on-off) loss of levodopa effects, and dyskinesia consists of hyperkinetic involuntary movements triggered by the drug. Both phenomena affect 40% of patients after 4–6 years of treatment.3, 4 Visual hallucinations’ estimated prevalence is about 22–38% of PD patients who use dopaminergic therapy.2 These complications are phenotypic features that vary between patients, and their risk factors are not completely understood. The theoretical framework to explain this differential individual response to the drug incorporates at least two concepts: (1) disease-associated features, such as severity of nigrostriatal denervation, and (2) pharmacological factors, given the non-physiological dopaminergic stimulation provided by levodopa.5, 6, 7 However, a growing body of evidence points to aberrant neuroplasticity, along with glutamatergic hyperactivity, as a third concept in this model, especially for dyskinesia.8

HOMER1 encodes a postsynaptic density protein highly expressed in the brain. This protein has important roles in synaptic plasticity and glutamate signal transduction.9, 10 This gene was mapped at chromosome 5, and it has two major products determined by alternative splicing. Homer1b/c is the long variant constitutively expressed and forms dimers that connect glutamate receptors (mGluR1α and mGluR5) with signaling proteins (for example, Shank and IP3R), whereas Homer1a is the short variant of the gene transcript and is an immediate early gene product expressed under neuronal activation. Homer1a lacks the carboxy-terminal region, which prevents dimer formation and counteracts the action of Homer1b/c. Previous studies with parkinsonism animal models showed an overexpression of Homer1a after the administration of dopaminergic agonists11, 12, 13 and a decreased expression after high-frequency deep brain stimulation of the subthalamic nucleus, a similar kind of stimulation used in clinical practice to improve symptoms in PD patients.14 This evidence suggested a relationship between Homer1 and symptomatic therapies for PD.

The possibility of predicting complications of the chronic use of levodopa in PD is an important goal toward a more personalized prescription. The aim of this study was to determine the contribution of rs4704559, rs10942891 and rs4704560 polymorphisms in the HOMER1 promoter region in the occurrence of chronic complications of levodopa therapy, namely motor fluctuations, dyskinesia and visual hallucinations.

Patients and methods


A total of 205 Brazilian PD patients were evaluated at the Movement Disorder Clinics at ‘Hospital de Clínicas de Porto Alegre’, Brazil, from 2006 to 2012. This sample has been fully described elsewhere.15, 16 Briefly, diagnosis of idiopathic PD was made in accordance with UK Parkinson Disease Society Brain Bank criteria17 and revised by one of us (CRMR), an experienced neurologist in movement disorders. The inclusion criteria were the use of at least 200 mg of levodopa for 1 year or more with an initial good response to the drug. Patients with atypical manifestations or secondary parkinsonisms were excluded. The hospital Ethics Committee approved the study, and all participants gave written informed consent.

Clinical evaluation

Patients and their caregivers underwent a structured interview for collecting clinical and demographic data. Information was also obtained from an electronic database of medical records. A trained physician applied the four parts of the Unified Parkinson’s Disease Rating Scale (UPDRS),18 the Hoehn and Yahr (HY) Scale19 and the Mini Mental State Examination.20

The principal outcomes investigated were motor fluctuations, dyskinesia and visual hallucinations, and they were defined by the presence of at least one of two criteria. For motor fluctuation, the criteria were as follows: (1) a score of one or more in question 39 of the UPDRS and/or (2) intake of levodopa five or more times a day. For dyskinesia, the criteria were as follows: (1) Dyskinesia is a chronic adverse effect, and therefore it is not possible to determine the number of episodes; question 32 of the UPDRS evaluates how long during the day dyskinesia is present. A score of one (1–25% of the day) or more in question 32 of the UPDRS and/or (2) any change in medical treatment due to dyskinesia (for example, adjustment in dopaminergic therapy or use of amantadine due to dyskinesia). For visual hallucinations, the criteria were as follows: (1) a score of two or more in the second question of the UPDRS and/or (2) any change in medical treatment due to hallucinations (for example, adjustment in dopaminergic therapy or use of atypical antipsychotic due to hallucinations). Historical information regarding these outcomes was assessed directly with the patient and/or caregiver and also with a review of an electronic database of medical records at the Movement Disorder Clinics.


Genomic DNA was extracted from peripheral blood samples by standard procedures.21 Three single-nucleotide polymorphisms (rs4704559, rs10942891 and rs4704560) in the HOMER1 5′ flanking region were selected, on the basis of their minor allele frequency and a previous study.22 The polymorphisms were genotyped by allelic discrimination using TaqMan assays according to the manufacturer's recommended protocols (Applied Biosystems, Foster City, CA, USA).

Statistical analyses

Allele frequencies were estimated by gene counting. Agreement of genotype frequencies to Hardy–Weinberg expectations was tested using a goodness-of-fit χ2 test. The Haploview program V3.3223 was used to estimate linkage disequilibrium (LD;D′) as a measure of pairwise LD and to create a graphical representation of LD structure.24 Haplotype frequencies and odds ratios were estimated using the default full model implemented in UNPHASED software (version 3.1.7), controlling for covariates. It tests the null hypothesis that there is no difference in OR between haplotypes by using a likelihood-ratio test. P-values were estimated and adjusted for 1000 permutations. Individual P-values regarding the comparison between each haplotype and the reference one was performed using the UNPHASED ‘compare with’ option. Bonferroni correction was used for multiple comparisons when it was the case. All other analyses were performed using the SPSS version 18 software. Confounders were previously defined by statistical or conceptual criteria, on the basis of literature review.4, 5, 7 Statistical confounders were those associated with genetic markers and with the outcome at P0.1 in our sample. For categorical variables, association was verified with χ2 test, and for continuous data, with or without normal distribution, Student’s t test or Wilcoxon–Mann–Whitney was performed, respectively. Conceptual confounders for motor fluctuations were disease duration and severity of disease (assessed indirectly through HY Scale). For dyskinesia and visual hallucinations, age at onset, disease duration, HY Scale, use of dopamine agonist and levodopa dose corrected by the concomitant use of catechol-O-methyl transferase inhibitor were considered.25 Gender was also selected as a conceptual confounder for dyskinesia.26 European ancestry was included in all analyses in order to control for a possible effect of population stratification.

Multiple Poisson regression with robust variance estimators was used to assess the effect of polymorphisms on the occurrence of motor fluctuations, dyskinesia and visual hallucinations, controlling for putative confounders. This method provides correct estimates, and it is a better alternative for cross-sectional studies with binary outcomes with high frequencies (>10%) than logistic regression, as the prevalence ratio is more interpretable.27, 28 Statistical significance was defined as a two-tailed P-value <0.05.


Minor allele frequencies for each polymorphism are as follows: 0.18 for rs4704559, 0.47 for rs4704560 and 0.18 for rs10942891. Genotype distributions are in accordance with the Hardy–Weinberg equilibrium (data not shown). The examination of pairwise LD indicated that the three polymorphisms were in strong LD (D′ higher than 0.96; P<0.001 in all comparisons; Supplementary Table 1).

Clinical and demographic data are presented in Table 1. A total of 176 individuals (83.7%) were of exclusively European ancestry (Table 1), and 16.3% were of African ancestry. The distribution of ancestry according to the outcomes studied (motor fluctuation, dyskinesia and visual hallucination) was similar among groups. Patients were aged between 43 and 94 years (67.38±10.34 years), with a mean age at onset of 58.77±11.07 years, and 51.2% were men. Motor fluctuations were present in 58% of patients, and this group had an earlier age at onset of motor symptoms, longer disease duration and used higher levodopa doses. The use of dopamine agonists and HY was not associated with motor fluctuations. Dyskinesia was present in 42% of the patients and was not associated with gender. This group exhibited an earlier age at onset and longer disease duration. Dyskinesia was also associated with higher levodopa doses and higher frequency of dopamine agonist and catechol-O-methyl transferase inhibitor use. Fifty-four subjects (26.3%) presented with visual hallucinations, and men were more affected. Age at onset was earlier in this group, and disease duration and levodopa doses were higher. There were no associations between visual hallucinations and disease severity, concomitant use of dopamine agonists or catechol-O-methyl transferase inhibitors.

Table 1: Demographic and clinical data, overall and according to outcomes

Univariate analyses for individual polymorphisms in a dominant model are shown in Table 2. The HOMER1 rs4704559 G allele was associated with a lower prevalence of dyskinesia (P=0.043) and visual hallucinations (P=0.068). Motor fluctuations were not associated with this polymorphism. No significant associations were observed between the other single nucleotide polymorphisms and levodopa adverse effects.

Table 2: Univariate analyses for the effect of individual polymorphisms on the occurrence of motor fluctuation, dyskinesia and visual hallucination

Multiple Poisson regression analyses were performed for the association between the rs4704559 A/G polymorphism with the presence of dyskinesia and visual hallucinations, controlling for the conceptual confounders previously described. On the basis of our previous definition, there was no statistical confounder. Age at onset and disease duration were correlated (r=−0.37, P<0.001), and therefore only the first variable was included in final models to avoid multicollinearity. Age was dichotomized in 59 years because of a nonlinear effect of age at onset on visual hallucinations. The rs4704559 G allele was associated with lower frequencies of dyskinesia and visual hallucinations (dyskinesia: prevalence ratio=0.615, 95% confidence interval 0.426–0.887, P=0.009; visual hallucinations: prevalence ratio=0.515, 95% confidence interval 0.295–0.899, P=0.020; Tables 3 and 4).

Table 3: Multiple Poisson regression analyses of HOMER1 polymorphism rs4704559 predicting dyskinesia among Parkinson’s disease patients who use levodopa
Table 4: Multiple Poisson regression analyses of HOMER1 polymorphism rs4704559 predicting visual hallucination among Parkinson’s disease patients who use levodopa

Dyskinesia was also associated with increased disease severity (HY scale), higher levodopa dose corrected for catechol-O-methyl transferase inhibitor and lower age at onset of motor symptoms (Table 4), whereas visual hallucinations were associated with earlier age at onset of motor symptoms (Table 5). Potential interaction effects between the polymorphism and covariates were considered in both models without significant results (all P>0.4). The overall adequacy of the final models was estimated by the likelihood-ratio χ2 test (dyskinesia: P<0.001; visual hallucinations: P=0.025).

Table 5: Multivariate analysis using UNPHASED for the presence of the haplotype GCT and the presence of dyskinesia and visual hallucination

Haplotype analysis was performed for dyskinesia and visual hallucination, controlling for previously described covariates (Table 5). The GCT haplotype was compared with the other haplotypes, as this was the only one that had the rs4704559 G allele that was associated with the outcomes. The presence of the GCT haplotype was associated with dyskinesia (odds ratios=0.47, 95% confidence interval 0.26–0.87, adjusted P=0.021) and with visual hallucination (odds ratios=0.46, 95% confidence interval 0.24–0.91, adjusted P=0.019), as expected from the analysis of the single polymorphism (Tables 3 and 4).


This study evaluated candidate polymorphisms in the promoter region of HOMER1, a gene involved in glutamate transmission and neuroplasticity in relation to the adverse effects induced by chronic use of levodopa in Parkinson's disease patients. The major findings reported herein were the association of the rs4704559 G allele with a lower prevalence of levodopa-induced dyskinesia and visual hallucinations, suggesting that this allele has a protective role for the development of levodopa adverse effects.

Visual hallucinations were associated with a higher levodopa dose in our sample. This finding was previously reported in one study, but not all.7, 29 Age at onset of motor symptoms was lower in patients with visual hallucinations, but this fact could represent higher disease duration and more time on medication. In the same line of earlier studies,30, 31 our patients with hallucinations have worse scores in UPDRS II, which represents activities of daily living, and UPDRS III, which represents motor impairment.

A previous study reported an association between the HOMER1 rs4704559 A allele and higher risk of hallucinations in 131 PD patients.22 The present study is the second independent demonstration that HOMER1 rs4704559 influences the occurrence of hallucinations and may be one of the genetic risk factors for the occurrence of this levodopa side effect. Considering that the rare G allele of this polymorphism was associated with a lower prevalence of visual hallucinations in the present sample, our result could be regarded as a replication of that report in a larger sample with a different ethnic background.

Implication of glutamate in the origin of levodopa-induced dyskinesia is not a new concept and, in fact, the only drug approved to treat this condition is a glutamate antagonist—amantadine.32 The pulsatile stimulation provided by levodopa therapy leads to hyperactivity of the glutamate transmission, observed both in animal models and in humans with dyskinesia.33 Homer1 is a family of proteins involved in glutamatergic transmission, and the expression of its short variant (Homer1a) is induced by dopamine agonists.11, 12, 13 Despite these data, the relationship between levodopa-induced dyskinesia and HOMER1 polymorphisms has not been investigated so far. Previous reports mostly investigated the association with polymorphisms in genes involved in dopamine transmission.15, 16, 34, 35 We showed an association between the HOMER1 rs4704559 G allele and lower prevalence of levodopa-induced dyskinesia in this study, suggesting a protective effect of this allele for the development of this adverse effect.

There are no reports about the potential function of the rs4704559 polymorphism and the mechanisms through which it influences dyskinesia and/or visual hallucinations in levodopa-treated PD patients. Considering this lack of information, it could be hypothesized that this polymorphism, in the 5′ UTR, close to the gene promoter, has a role in transcription regulation. Homer1 has two variants: Homer1b/c, constitutively expressed, and Homer1a, expressed under special circumstances as an immediate early gene product.10 It has been demonstrated that dopamine agonists induced the expression of Homer1a; as the rs4704559 polymorphism is close to the promoter region, it could potentially affect the expression of this induced variant.11, 12, 13 Therefore, it is possible that it interferes with the formation of Homer1 dimers, disrupts the link between the glutamate receptor and its second messengers and affects glutamatergic transmission. Clearly, more studies are needed to disclose the functional effect of this variant on gene function.

The overall results presented in this study should be viewed in the context of some limitations. The sample is of moderate size, and thus these data require additional confirmation in a larger sample. Because this study is cross-sectional, it was not possible to trace the exact time that each patient takes to develop dyskinesia. A longitudinal design would be preferable to better delineate the time at onset of adverse effects and to better appreciate the importance of each covariate as risk factors. Although a study with 48 ancestry informative markers did not show significant population structure in Southern Brazilians,36 no genomic control was performed in the present work. Therefore, our findings could have been biased by hidden genetic heterogeneity, but we used ethnicity as a covariate in all analyses. The study design is observational-naturalistic, which is valuable to better appreciate the role of genetic factors in routine clinical practice beyond the realm of controlled clinical trials. Despite this, in a naturalistic design, the controlling of confounders is a challenge. However, it is unlikely that the results could be attributed to other factors, because an extensive assessment of potential confounders between groups was performed. Moreover, the previously described association between the rs4704559 polymorphism and hallucinations was also observed in a naturalistic study.22

PD, between all the neurodegenerative disorders, is the only one that has an important symptomatic therapy. However, these treatments can lead to adverse events that have an impact on the quality of life and are a challenge to the clinician. Understanding the mechanisms underlying this variability in pharmacological response would be an important tool to choose better treatment strategies, to better select patients for clinical trials and to give new insights for drug discovery. Although there is growing evidence linking glutamate transmission and neuroplasticity with PD, few attempts were made to investigate the relationship of genes associated with these processes and the response to dopaminergic drugs. The results presented herein provide evidence for a relationship between HOMER1 and adverse effects induced by levodopa therapy. Future neuroimaging and pathological studies including genes will help develop new diagnostic tools and more personalized therapies.


  1. 1.

    , . Epidemiology of Parkinson’s disease. Lancet Neurology 2006; 5: 525–535.

  2. 2.

    , . Epidemiology of psychosis in Parkinson’s disease. J Neurol Sci 2010; 289: 12–17.

  3. 3.

    , . Frequency of levodopa-related dyskinesias and motor fluctuations as estimated from the cumulative literature. Mov Disord 2001; 16: 448–458.

  4. 4.

    . Motor fluctuations and dyskinesia in Parkinson’s disease. Mov Disord 2001; 8: 101–108.

  5. 5.

    , , . Factors associated with the development of motor fluctuations and dyskinesias in Parkinson disease. Arch Neurol 2006; 63: 1756–1760.

  6. 6.

    , , , , . Levodopa-induced dyskinesias in patients with Parkinson’s disease: filling the bench-to-bedside gap. Lancet Neurol 2010; 9: 1106–1117.

  7. 7.

    , , , , , et al. A 12-year population-based study of psychosis in Parkinson disease. Arch Neurol 2010; 67: 996–1001.

  8. 8.

    . New ideas on the origin of L-dopa-induced dyskinesias: age, genes and neural plasticity. Trends Pharmacol Sci 2005; 26: 391–397.

  9. 9.

    . Modulation of synaptic signalling complexes by Homer proteins. J Neurochem 2002; 81: 407–413.

  10. 10.

    , , , . Scaffold protein Homer 1: implications for neurological diseases. Neurochem Int 2012; 61: 731–738.

  11. 11.

    , , , , . A complex program of striatal gene expression induced by dopaminergic stimulation. J Neurosci 1998; 18: 5301–5310.

  12. 12.

    , , , , , . Coordinated and spatial upregulation of arc in striatonigral neurons correlates with L-dopa-induced behavioral sensitization in dyskinetic rats. J Neuropathol Exp Neurol 2005; 64: 936–947.

  13. 13.

    , , , , , et al. The dopamine D1 receptor agonist, but not the D2 receptor agonist, induces gene expression of Homer 1a in rat striatum and nucleus accumbens. Brain Res 2007; 1131: 88–96.

  14. 14.

    , , , , , et al. Deep brain stimulation in a rat model modulates TH, CaMKIIa and Homer1 gene expression. Eur J Neurosci 2007; 25: 239–250.

  15. 15.

    , , , , , et al. DRD2 haplotype is associated with dyskinesia induced by levodopa therapy in Parkinson’s disease patients. Pharmacogenomics 2012; 13: 1701–1710.

  16. 16.

    , , , , , et al. Polymorphisms in the dopamine transporter gene are associated with visual hallucinations and levodopa equivalent dose in Brazilians with Parkinson’s disease. Int J Neuropsychopharmacol 2013; 16: 1251–1258.

  17. 17.

    , , , . What features improve the accuracy of clinical diagnosis in Parkinson’s disease: a clinicopathologic study. Neurology 1992; 42: 1142–1146.

  18. 18.

    , , , , , . Unified Parkinson’s Disease Rating Scale characteristics and structure. The Cooperative Multicentric Group. Mov Disord 1994; 9: 76–83.

  19. 19.

    , . Parkinsonism: onset, progression and mortality. Neurology 1967; 17: 427–442.

  20. 20.

    , , . ‘Mini-mental state’. A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975; 12: 189–198.

  21. 21.

    , , , , . A non-organic and non-enzymatic extraction method gives higher yields of genomic DNA from whole-blood samples than do nine other methods tested. J Biochem Biophys Methods 1992; 25: 193–205.

  22. 22.

    , , , , , et al. HOMER1 promoter analysis in Parkinson’s disease: association study with psychotic symptoms. Neuropsychobiology 2009; 59: 239–245.

  23. 23.

    . Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Hum Hered 2008; 66: 89–98.

  24. 24.

    , , , . Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics 2005; 21: 263–265.

  25. 25.

    , , , , , . Systematic review of levodopa dose equivalency reporting in Parkinson’s disease. Mov Disord 2010; 25: 2649–2653.

  26. 26.

    , , , , , et al. Sex differences in clinical and genetic determinants of levodopa peak-dose dyskinesias in Parkinson disease: an exploratory study. Arch Neurol 2005; 62: 601–605.

  27. 27.

    . The (relative) risks of using odds ratios. Arch Dermatol 2006; 142: 761–764.

  28. 28.

    , . Alternatives for logistic regression in cross-sectional studies: an empirical comparison of models that directly estimate the prevalence ratio. BMC Med Res Methodol 2003; 3: 21.

  29. 29.

    , , , , . Prospective longitudinal assessment of hallucinations in Parkinson’s disease. Neurology 2001; 57: 2078–2082.

  30. 30.

    , , . Prospective study of hallucinations and delusions in Parkinson’s disease. J Neurol Neurosurg Psychiatr 2001; 70: 734–738.

  31. 31.

    , , , . Prevalence and clinical correlates of psychotic symptoms in Parkinson disease: a community-based study. Arch Neurol 1999; 56: 595–601.

  32. 32.

    , , . N-Methyl-D-Aspartate antagonists in levodopa induced dyskinesia: a meta-analysis. Can J Neurol Sci 2012; 39: 465–472.

  33. 33.

    , , , , , et al. Glutamate NMDA receptor dysregulation in Parkinson’s disease with dyskinesias. Brain 2011; 134: 979–986.

  34. 34.

    , , , , , . The association of functional catechol-O-methyl transferase haplotypes with risk of Parkinson’s disease, levodopa treatment response, and complications. Pharmacogenet Genomics 2008; 18: 815–821.

  35. 35.

    , , , , , . Motor complications in patients form the German Competence Network on Parkinson’s disease and the DRD3 Ser9Gly polymorphism. Mov Disord 2009; 24: 1080–1084.

  36. 36.

    , , , , , et al. Assessing individual interethnic admixture and population substructure using a 48-insertion-deletion (INSEL) ancestry-informative marker (AIM) panel. Hum Mutat 2010; 31: 184–190.

Download references


The research was supported by grants from ‘Financiadora de Estudos e Projetos’ (FINEP 01.08.01230.00), ‘Conselho Nacional de Desenvolvimento Científico e Tecnológico’ (CNPq) and ‘Fundo de Incentivo à Pesquisa (FIPE – HCPA). Source of grants: Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil), Financiadora de Estudos e Projetos (FINEP, Brazil) and Fundo de Incentivo à Pesquisa (FIPE – HCPA).

Author information


  1. Departamento de Genética, Instituto de Biociências, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil

    • A F Schumacher-Schuh
    • , V Altmann
    • , M Rieck
    • , L Tovo-Rodrigues
    •  & M H Hutz
  2. Serviço de Neurologia, Hospital de Clínicas de Porto Alegre, Porto Alegre, Brazil

    • A F Schumacher-Schuh
    • , T L Monte
    • , M S Medeiros
    •  & C R M Rieder
  3. Departamento de Estatística, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil

    • S M Callegari-Jacques


  1. Search for A F Schumacher-Schuh in:

  2. Search for V Altmann in:

  3. Search for M Rieck in:

  4. Search for L Tovo-Rodrigues in:

  5. Search for T L Monte in:

  6. Search for S M Callegari-Jacques in:

  7. Search for M S Medeiros in:

  8. Search for C R M Rieder in:

  9. Search for M H Hutz in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to M H Hutz.

Supplementary information

About this article

Publication history







Supplementary Information accompanies the paper on the The Pharmacogenomics Journal website (http://www.nature.com/tpj)

Further reading