Transcript-aware analysis of rare predicted loss-of-function variants in the UK Biobank elucidate new isoform-trait associations.

A single gene can produce multiple transcripts with distinct molecular functions. Rare-variant association tests often aggregate all coding variants across individual genes, without accounting for the variants' presence or consequence in resulting transcript isoforms. To evaluate the utility of transcript-aware variant sets, rare predicted loss-of-function (pLOF) variants were aggregated for 17,035 protein-coding genes using 55,558 distinct transcript-specific variant sets. These sets were tested for their association with 728 circulating proteins and 188 quantitative phenotypes across 406,921 individuals in the UK Biobank. The transcript-specific approach resulted in larger estimated effects of pLOF variants decreasing serum cis-protein levels compared to the gene-based approach (pbinom ≤ 2x10-16). Additionally, 251 quantitative trait associations were identified as being significant using the transcript-specific approach but not the gene-based approach, including PCSK5 transcript ENST00000376752 and standing height (transcript-specific statistic, P = 1.3x10-16, effect = 0.7 SD decrease; gene-based statistic, P = 0.02, effect = 0.05 SD decrease) and LDLR transcript ENST00000252444 and apolipoprotein B (transcript-specific statistic, P = 5.7x10-20, effect = 1.0 SD increase; gene-based statistic, P = 3.0x10-4, effect = 0.2 SD increase). This approach demonstrates the importance of considering the effect of pLOFs on specific transcript isoforms when performing rare-variant association studies.

Dopamine pathway and Parkinson's risk variants are associated with levodopa-induced dyskinesia.

Background: Levodopa-induced dyskinesia (LID) is a common adverse effect of levodopa, one of the main therapeutics used to treat the motor symptoms of Parkinson's disease (PD). Previous evidence suggests a connection between LID and a disruption of the dopaminergic system as well as genes implicated in PD, including GBA1 and LRRK2. Objectives: To investigate the effects of genetic variants on risk and time to LID. Methods: We performed a genome-wide association study (GWAS) and analyses focused on GBA1 and LRRK2 variants. We also calculated polygenic risk scores including risk variants for PD and variants in genes involved in the dopaminergic transmission pathway. To test the influence of genetics on LID risk we used logistic regression, and to examine its impact on time to LID we performed Cox regression including 1,612 PD patients with and 3,175 without LID. Results: We found that GBA1 variants were associated with LID risk (OR=1.65, 95% CI=1.21-2.26, p=0.0017) and LRRK2 variants with reduced time to LID onset (HR=1.42, 95% CI=1.09-1.84, p=0.0098). The fourth quartile of the PD PRS was associated with increased LID risk (ORfourth_quartile=1.27, 95% CI=1.03-1.56, p=0.0210). The third and fourth dopamine pathway PRS quartiles were associated with a reduced time to development of LID (HRthird_quartile=1.38, 95% CI=1.07-1.79, p=0.0128; HRfourth_quartile=1.38, 95% CI=1.06-1.78, p=0.0147). Conclusions: This study suggests that variants implicated in PD and in the dopaminergic transmission pathway play a role in the risk/time to develop LID. Further studies will be necessary to examine how these findings can inform clinical care.

HLA in isolated REM sleep behavior disorder and Lewy body dementia.

Synucleinopathies-related disorders such as Lewy body dementia (LBD) and isolated/idiopathic REM sleep behavior disorder (iRBD) have been associated with neuroinflammation. In this study, we examined whether the human leukocyte antigen (HLA) locus plays a role in iRBD and LBD. In iRBD, HLA-DRB1*11:01 was the only allele passing FDR correction (OR = 1.57, 95% CI = 1.27-1.93, p = 2.70e-05). We also discovered associations between iRBD and HLA-DRB1 70D (OR = 1.26, 95%CI = 1.12-1.41, p = 8.76e-05), 70Q (OR = 0.81, 95%CI = 0.72-0.91, p = 3.65e-04) and 71R (OR = 1.21, 95%CI = 1.08-1.35, p = 1.35e-03). Position 71 (pomnibus = 0.00102) and 70 (pomnibus = 0.00125) were associated with iRBD. Our results suggest that the HLA locus may have different roles across synucleinopathies.

Human gain-of-function variants in HNF1A confer protection from diabetes but independently increase hepatic secretion of atherogenic lipoproteins.

Loss-of-function mutations in hepatocyte nuclear factor 1A (HNF1A) are known to cause rare forms of diabetes and alter hepatic physiology through unclear mechanisms. In the general population, 1:100 individuals carry a rare, protein-coding HNF1A variant, most of unknown functional consequence. To characterize the full allelic series, we performed deep mutational scanning of 11,970 protein-coding HNF1A variants in human hepatocytes and clinical correlation with 553,246 exome-sequenced individuals. Surprisingly, we found that ∼1:5 rare protein-coding HNF1A variants in the general population cause molecular gain of function (GOF), increasing the transcriptional activity of HNF1A by up to 50% and conferring protection from type 2 diabetes (odds ratio [OR] = 0.77, p = 0.007). Increased hepatic expression of HNF1A promoted a pro-atherogenic serum profile mediated in part by enhanced transcription of risk genes including ANGPTL3 and PCSK9. In summary, ∼1:300 individuals carry a GOF variant in HNF1A that protects carriers from diabetes but enhances hepatic secretion of atherogenic lipoproteins.

Genome-wide association study stratified by MAPT haplotypes identifies potential novel loci in Parkinson's disease.

Objective: To identify genetic factors that may modify the effects of the MAPT locus in Parkinson's disease (PD). Methods: We used data from the International Parkinson's Disease Genomics Consortium (IPDGC) and the UK biobank (UKBB). We stratified the IPDGC cohort for carriers of the H1/H1 genotype (PD patients n=8,492 and controls n=6,765) and carriers of the H2 haplotype (with either H1/H2 or H2/H2 genotypes, patients n=4,779 and controls n=4,849) to perform genome-wide association studies (GWASs). Then, we performed replication analyses in the UKBB data. To study the association of rare variants in the new nominated genes, we performed burden analyses in two cohorts (Accelerating Medicines Partnership - Parkinson Disease and UKBB) with a total sample size PD patients n=2,943 and controls n=18,486. Results: We identified a novel locus associated with PD among MAPT H1/H1 carriers near EMP1 (rs56312722, OR=0.88, 95%CI= 0.84-0.92, p= 1.80E-08), and a novel locus associated with PD among MAPT H2 carriers near VANGL1 (rs11590278, OR=1.69 95%CI=1.40-2.03, p=2.72E-08). Similar analysis of the UKBB data did not replicate these results and rs11590278 near VANGL1 did have similar effect size and direction in carriers of H2 haplotype, albeit not statistically significant (OR= 1.32, 95%CI= 0.94-1.86, p=0.17). Rare EMP1 variants with high CADD scores were associated with PD in the MAPT H2 stratified analysis (p=9.46E-05), mainly driven by the p.V11G variant. Interpretation: We identified several loci potentially associated with PD stratified by MAPT haplotype and larger replication studies are required to confirm these associations.

NPC1 variants are not associated with Parkinson's disease, REM-sleep behavior disorder or dementia with Lewy bodies in European cohorts.

NPC1 encodes a lysosomal protein involved in cholesterol transport. Biallelic mutations in this gene may lead to Niemann-Pick disease type C (NPC), a lysosomal storage disorder. The role of NPC1 in alpha synucleinopathies is still unclear, as different genetic, clinical, and pathological studies have reported contradictory results. This study aimed to evaluate the association of NPC1 variants with the synucleinopathies Parkinson's disease (PD), dementia with Lewy bodies (DLB), and rapid eye movement-sleep behavior disorder (RBD). We analyzed common and rare variants from 3 cohorts of European descent: 1084 RBD cases and 2945 controls, 2852 PD cases and 1686 controls, and 2610 DLB cases and 1920 controls. Logistic regression models were used to assess common variants while optimal sequence Kernel association tests were used to assess rare variants, both adjusted for sex, age, and principal components. No variants were associated with any of the synucleinopathies, supporting that common and rare NPC1 variants do not play an important role in alpha synucleinopathies.

HLA in isolated REM sleep behavior disorder and Lewy body dementia.

Background and Objectives: Isolated/idiopathic REM sleep behavior disorder (iRBD) and Lewy body dementia (LBD) are synucleinopathies that have partial genetic overlap with Parkinson's disease (PD). Previous studies have shown that neuroinflammation plays a substantial role in these disorders. In PD, specific residues of the human leukocyte antigen ( HLA ) were suggested to be associated with a protective effect. This study examined whether the HLA locus plays a similar role in iRBD, LBD and PD. Methods: We performed HLA imputation on iRBD genotyping data (1,072 patients and 9,505 controls) and LBD whole-genome sequencing (2,604 patients and 4,032 controls) using the multi-ethnic HLA reference panel v2 from the Michigan Imputation Server. Using logistic regression, we tested the association of HLA alleles, amino acids and haplotypes with disease susceptibility. We included age, sex and the top 10 principal components as covariates. We also performed an omnibus test to examine which HLA residue positions explain the most variance. Results: In iRBD, HLA-DRB1 *11:01 was the only allele passing FDR correction (OR=1.57, 95% CI=1.27-1.93, p =2.70e-05). We also discovered associations between iRBD and HLA-DRB1 70D (OR=1.26, 95%CI=1.12-1.41, p =8.76e-05), 70Q (OR=0.81, 95% CI=0.72-0.91, p =3.65e-04) and 71R (OR=1.21, 95% CI=1.08-1.35, p =1.35e-03). In HLA-DRB1 , position 71 ( p omnibus =0.00102) and 70 ( p omnibus =0.00125) were associated with iRBD. We found no association in LBD. Discussion: This study identified an association between HLA-DRB1 11:01 and iRBD, distinct from the previously reported association in PD. Therefore, the HLA locus may play different roles across synucleinopathies. Additional studies are required better to understand HLA's role in iRBD and LBD.

Genome-wide association study of REM sleep behavior disorder identifies polygenic risk and brain expression effects.

Rapid-eye movement (REM) sleep behavior disorder (RBD), enactment of dreams during REM sleep, is an early clinical symptom of alpha-synucleinopathies and defines a more severe subtype. The genetic background of RBD and its underlying mechanisms are not well understood. Here, we perform a genome-wide association study of RBD, identifying five RBD risk loci near SNCA, GBA, TMEM175, INPP5F, and SCARB2. Expression analyses highlight SNCA-AS1 and potentially SCARB2 differential expression in different brain regions in RBD, with SNCA-AS1 further supported by colocalization analyses. Polygenic risk score, pathway analysis, and genetic correlations provide further insights into RBD genetics, highlighting RBD as a unique alpha-synucleinopathy subpopulation that will allow future early intervention.

Gene-based burden analysis of damaging private variants in PRKN, PARK7 and PINK1 in Parkinson's disease cohorts of European descent.

Recessive mutations in PRKN, PARK7, and PINK1 are established causes of early-onset Parkinson's disease (EOPD). Previous studies have interrogated the role of heterozygous variants in these genes but mainly focused on rare (minor allele frequency [MAF] <1%) damaging variants or established mutations. Here, we assessed heterozygous private PRKN, PARK7 and PINK1 variants in PD risk in four large-scale PD case-control datasets by performing gene-wise burden analyses using sequencing data totaling 5,829 PD cases and 7,221 controls, and summary allele counts from 9,501 PD cases and 48,207 controls. Results showed no significant burden in all three genes after meta-analyses. Burden in EOPD (age at onset <50 years) and late-onset PD (≥50 years) remained nonsignificant. In summary, we found no evidence to support the association of the excess burden of heterozygous private variants in PRKN, PARK7, and PINK1 with PD risk in the European population. Larger, more diverse cohorts are needed to accurately determine their role in PD.

Rapid eye movement sleep behaviour disorder: Past, present, and future.

This manuscript presents an overview of REM sleep behaviour disorder (RBD) with a special focus on European contributions. After an introduction examining the history of the disorder, we address the pathophysiological and clinical aspects, as well as the diagnostic issues. Further, implications of RBD diagnosis and biomarkers are discussed. Contributions of European researchers to this field are highlighted.

Rare PSAP Variants and Possible Interaction with GBA in REM Sleep Behavior Disorder.

BACKGROUND: PSAP encodes saposin C, the co-activator of glucocerebrosidase, encoded by GBA. GBA mutations are associated with idiopathic/isolated REM sleep behavior disorder (iRBD), a prodromal stage of synucleinopathy. OBJECTIVE: To examine the role of PSAP mutations in iRBD. METHODS: We fully sequenced PSAP and performed Optimized Sequence Kernel Association Test in 1,113 iRBD patients and 2,324 controls. We identified loss-of-function (LoF) mutations, which are very rare in PSAP, in three iRBD patients and none in controls (uncorrected p = 0.018). RESULTS: Two variants were stop mutations, p.Gln260Ter and p.Glu166Ter, and one was an in-frame deletion, p.332_333del. All three mutations have a deleterious effect on saposin C, based on in silico analysis. In addition, the two carriers of p.Glu166Ter and p.332_333del mutations also carried a GBA variant, p.Arg349Ter and p.Glu326Lys, respectively. The co-occurrence of these extremely rare PSAP LoF mutations in two (0.2%) GBA variant carriers in the iRBD cohort, is unlikely to occur by chance (estimated co-occurrence in the general population based on gnomAD data is 0.00035%). Although none of the three iRBD patients with PSAP LoF mutations have phenoconverted to an overt synucleinopathy at their last follow-up, all manifested initial signs suggestive of motor dysfunction, two were diagnosed with mild cognitive impairment and all showed prodromal clinical markers other than RBD. Their probability of prodromal PD, according to the Movement Disorder Society research criteria, was 98% or more. CONCLUSION: These results suggest a possible role of PSAP variants in iRBD and potential genetic interaction with GBA, which requires additional studies.

Fine mapping of the HLA locus in Parkinson's disease in Europeans.

We fine mapped the leukocyte antigen (HLA) region in 13,770 Parkinson's disease (PD) patients, 20,214 proxy-cases, and 490,861 controls of European origin. Four HLA types were associated with PD after correction for multiple comparisons, HLA-DQA1*03:01, HLA-DQB1*03:02, HLA-DRB1*04:01, and HLA-DRB1*04:04. Haplotype analyses followed by amino acid analysis and conditional analyses suggested that the association is protective and primarily driven by three specific amino acid polymorphisms present in most HLA-DRB1*04 subtypes-11V, 13H, and 33H (OR = 0.87, 95% CI: 0.83-0.90, p < 8.23 × 10-9 for all three variants). No other effects were present after adjustment for these amino acids. Our results suggest that specific HLA-DRB1 variants are associated with reduced risk of PD, providing additional evidence for the role of the immune system in PD. Although effect size is small and has no diagnostic significance, understanding the mechanism underlying this association may lead to the identification of new targets for therapeutics development.

Analysis of PTRHD1 common and rare variants in European patients with Parkinson's disease.

Three studies in Iranian and African families identified three different variants in the peptidyl-tRNA hydrolase domain containing 1 gene (PTRHD1) in patients affected by parkinsonism with intellectual impairment. In the current study, our objective was to investigate whether PTRHD1 variants are associated with Parkinson's disease (PD) risk and age at onset (AAO). To evaluate the association between PTRHD1 and PD risk, we analyzed whole genome sequencing data of 1647 PD cases and 1050 healthy controls, as well as genome-wide imputed genotyping data on 14,671 PD cases and 17,667 controls, all of European ancestry. Furthermore, we examined the association of PTRHD1 with PD risk and AAO using summary statistics data from the most recent PD genome-wide association study meta-analyses. Our results show no association between PTRHD1 and PD risk or AAO. We conclude that PTRHD1 does not play a major role in PD in the European population. Further large-scale studies including subjects with different ancestry and family trios are required.

Lack of Causal Effects or Genetic Correlation between Restless Legs Syndrome and Parkinson's Disease.

BACKGROUND: Epidemiological studies have reported an association between Parkinson's disease (PD) and restless legs syndrome. OBJECTIVES: We aimed to use genetic data to study whether these 2 disorders are causally linked or share genetic architecture. METHODS: We performed two-sample Mendelian randomization and linkage disequilibrium score regression using summary statistics from recent genome-wide meta-analyses of PD and restless legs syndrome. RESULTS: We found no evidence for a causal relationship between restless legs syndrome (as the exposure) and PD (as the outcome, inverse variance-weighted; b = -0.003, SE = 0.031, P = 0.916; F statistic = 217.5). Reverse Mendelian randomization also did not demonstrate any causal effect of PD on restless legs syndrome (inverse variance-weighted; b = -0.012, SE = 0.023, P = 0.592; F statistic = 191.7). Linkage disequilibrium score regression analysis demonstrated lack of genetic correlation between restless legs syndrome and PD (rg = -0.028, SE = 0.042, P = 0.507). CONCLUSIONS: There was no evidence for a causal relationship or genetic correlation between restless legs syndrome and PD. The associations observed in epidemiological studies could be attributed, in part, to confounding or nongenetic determinants. © 2021 International Parkinson and Movement Disorder Society.

Investigation of Autosomal Genetic Sex Differences in Parkinson's Disease.

OBJECTIVE: Parkinson's disease (PD) is a complex neurodegenerative disorder. Men are on average ~ 1.5 times more likely to develop PD compared to women with European ancestry. Over the years, genomewide association studies (GWAS) have identified numerous genetic risk factors for PD, however, it is unclear whether genetics contribute to disease etiology in a sex-specific manner. METHODS: In an effort to study sex-specific genetic factors associated with PD, we explored 2 large genetic datasets from the International Parkinson's Disease Genomics Consortium and the UK Biobank consisting of 13,020 male PD cases, 7,936 paternal proxy cases, 89,660 male controls, 7,947 female PD cases, 5,473 maternal proxy cases, and 90,662 female controls. We performed GWAS meta-analyses to identify distinct patterns of genetic risk contributing to disease in male versus female PD cases. RESULTS: In total, 19 genomewide significant regions were identified and no sex-specific effects were observed. A high genetic correlation between the male and female PD GWAS were identified (rg = 0.877) and heritability estimates were identical between male and female PD cases (~ 20%). INTERPRETATION: We did not detect any significant genetic differences between male or female PD cases. Our study does not support the notion that common genetic variation on the autosomes could explain the difference in prevalence of PD between males and females cases at least when considering the current sample size under study. Further studies are warranted to investigate the genetic architecture of PD explained by X and Y chromosomes and further evaluate environmental effects that could potentially contribute to PD etiology in male versus female patients. ANN NEUROL 2021;90:41-48.

Assessment of ANG variants in Parkinson's disease.

Genetic risk factors are occasionally shared between different neurodegenerative diseases. Previous studies have linked ANG, a gene encoding angiogenin, to both Parkinson's disease (PD) and amyotrophic lateral sclerosis (ALS). Functional studies suggest ANG plays a neuroprotective role in both PD and amyotrophic lateral sclerosis by reducing cell death. We further explored the genetic association between ANG and PD by analyzing genotype data from the International Parkinson's Disease Genomics Consortium (14,671 cases and 17,667 controls) and whole genome sequencing data from the Accelerating Medicines Partnership - Parkinson's disease initiative (AMP-PD, https://amp-pd.org/) (1,647 cases and 1,050 controls). Our analysis did not replicate the findings of previous studies and identified no significant association between ANG variants and PD risk.

LRRK2 p.M1646T is associated with glucocerebrosidase activity and with Parkinson's disease.

The LRRK2 p.G2019S Parkinson's disease (PD) variant is associated with elevated glucocerebrosidase (GCase) activity in peripheral blood. We aimed to evaluate the association of other LRRK2 variants with PD and its association with GCase activity. LRRK2 and GBA were fully sequenced in 1123 PD patients and 576 controls from the Columbia and PPMI cohorts, in which GCase activity was measured in dried blood spots by liquid chromatography-tandem mass spectrometry. LRRK2 p.M1646T was associated with increased GCase activity in the Columbia University cohort (ß = 1.58, p = 0.0003), and increased but not significantly in the PPMI cohort (ß = 0.29, p = 0.58). p.M1646T was associated with PD (odds ratio = 1.18, 95% confidence interval = 1.09-1.28, p = 7.33E-05) in 56,306 PD patients and proxy-cases, and 1.4 million controls. Our results suggest that the p.M1646T variant is associated with risk of PD with a small effect and with increased GCase activity in peripheral blood.

Cytokines and Gaucher Biomarkers in Glucocerebrosidase Carriers with and Without Parkinson Disease.

BACKGROUND: Homozygous and compound heterozygous variants in glucocerebrosidase (GBA) can cause Gaucher disease (GD), whereas heterozygous variants increase the risk of developing Parkinson's disease (PD). GD patients display altered peripheral immune proteins. However, it is unknown if these are altered in GBA carriers with PD. OBJECTIVES: To determine whether plasma cytokines and immune biomarkers associated with GD are also altered in GBA carriers with or without PD. METHODS: Inflammatory cytokines and established GD biomarkers, ferritin, CD162, CCL18, and chitotriosidase (28 biomarkers) were measured in GBA pathogenic variant carriers with (n = 135) and without (n = 83) PD, and non-carriers with (n = 75) and without PD (n = 77). RESULTS: PD patients with biallelic pathogenic variants in GBA had elevated plasma levels of ferritin, CCL18, and MIP1α. These biomarkers were not elevated in heterozygous GBA carriers. CONCLUSION: GD plasma biomarkers are not promising candidates for stratifying the risk for PD in carriers of heterozygous GBA pathogenic variants. © 2021 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Common X-Chromosome Variants Are Associated with Parkinson Disease Risk.

OBJECTIVE: The objective of this study was to identify genetic variants on the X-chromosome associated with Parkinson disease (PD) risk. METHODS: We performed an X-chromosome-wide association study (XWAS) of PD risk by meta-analyzing results from sex-stratified analyses. To avoid spurious associations, we designed a specific harmonization pipeline for the X-chromosome and focused on a European ancestry sample. We included 11,142 cases, 280,164 controls, and 5,379 proxy cases, based on parental history of PD. Additionally, we tested the association of significant variants with (1) PD risk in an independent replication with 1,561 cases and 2,465 controls and (2) putamen volume in 33,360 individuals from the UK Biobank. RESULTS: In the discovery meta-analysis, we identified rs7066890 (odds ratio [OR] = 1.10, 95% confidence interval [CI] = 1.06-1.14, p = 2.2 × 10-9 ), intron of GPM6B, and rs28602900 (OR = 1.10, 95% CI = 1.07-1.14, p = 1.6 × 10-8 ) in a high gene density region including RPL10, ATP6A1, FAM50A, and PLXNA3. The rs28602900 association with PD was replicated (OR = 1.16, 95% CI = 1.03-1.30, p = 0.016) and shown to colocalize with a significant expression quantitative locus (eQTL) regulating RPL10 expression in the putamen and other brain tissues in the Genotype-Tissue Expression Project. Additionally, the rs28602900 locus was found to be associated with reduced brain putamen volume. No results reached genome-wide significance in the sex-stratified analyses. INTERPRETATION: We report the first XWAS of PD and identify 2 genome-wide significant loci. The rs28602900 association was replicated in an independent PD dataset and showed concordant effects in its association with putamen volume. Critically, rs26802900 is a significant eQTL of RPL10. These results support a role for ribosomal proteins in PD pathogenesis and show that the X-chromosome contributes to PD genetic risk. ANN NEUROL 2021;90:22-34.

Heritability Enrichment Implicates Microglia in Parkinson's Disease Pathogenesis.

OBJECTIVE: Understanding how different parts of the immune system contribute to pathogenesis in Parkinson's disease is a burning challenge with important therapeutic implications. We studied enrichment of common variant heritability for Parkinson's disease stratified by immune and brain cell types. METHODS: We used summary statistics from the most recent meta-analysis of genomewide association studies in Parkinson's disease and partitioned heritability using linkage disequilibrium score regression, stratified for specific cell types, as defined by open chromatin regions. We also validated enrichment results using a polygenic risk score approach and intersected disease-associated variants with epigenetic data and expression quantitative loci to nominate and explore a putative microglial locus. RESULTS: We found significant enrichment of Parkinson's disease risk heritability in open chromatin regions of microglia and monocytes. Genomic annotations overlapped substantially between these 2 cell types, and only the enrichment signal for microglia remained significant in a joint model. We present evidence suggesting P2RY12, a key microglial gene and target for the antithrombotic agent clopidogrel, as the likely driver of a significant Parkinson's disease association signal on chromosome 3. INTERPRETATION: Our results provide further support for the importance of immune mechanisms in Parkinson's disease pathogenesis, highlight microglial dysregulation as a contributing etiological factor, and nominate a targetable microglial gene candidate as a pathogenic player. Immune processes can be modulated by therapy, with potentially important clinical implications for future treatment in Parkinson's disease. ANN NEUROL 2021;89:942-951.