Which Of The Following Services Falls Closest To The Middle Of The Tangibility Continuum

Abstract

In only over a decade, advances in genome-wide association studies (GWAS) accept offered an approach to stratify individuals based on genetic risk for disease. Using recent Alzheimer's disease (Ad) GWAS results as the base of operations information, we adamant each private'south polygenic risk score (PRS) in the UK Biobank dataset. Using individuals within the extreme risk distribution, we performed a GWAS that is agnostic of AD phenotype and is instead based on known genetic risk for affliction. To interpret the functions of the new risk factors, we conducted phenotype analyses, including a phenome-wide clan study. Nosotros identified 246 loci surpassing the significance threshold of which 229 were not reported in the base AD GWAS. These include loci that showed suggestive levels of clan in the base GWAS and loci non previously suspected to be associated with Advert. Amid these, there are loci, such as IL34 and KANSL1, that have since been shown to be associated with Advertisement in contempo studies. We also bear witness highly significant genetic correlations with multiple health-related outcomes that provide insights into prodromal symptoms and comorbidities. This is the first study to utilize PRS equally a phenotype-agnostic group classification in Advertisement genetic studies. We identify potential new loci for Advertizing and detail phenotypic analysis of these PRS extremes.

Introduction

Alzheimer's disease (AD) is i of the almost mutual, disabling neurodegenerative diseases faced past our society¹. Heritability estimates from twin studies range from lx to eighty%^two, suggesting a strong genetic component to the disease. Even so, a meaning fraction of the phenotypic variance of the disease is unexplained by the currently known genome-wide significant loci³. Over the concluding decade, increasing sample sizes in Advertizement genome-wide association studies (GWAS) take greatly improved the statistical power to observe novel genetic associations^4,5,6,7,8. In addition, recent studies have characterized novel rare variability in the illness, furthering our agreement of genetic mechanisms underlying AD⁹.

Although increasing sample size is a tested approach to place new loci in complex disease inquiry, innovative approaches to further investigate these in big datasets may harbor further insights into the currently missing heritability.

Polygenic gamble scores (PRS) have been used to empathize the genetic liability of developing specific traits¹⁰. PRS are calculated from a set of independent variants associated with the disease or trait under study¹¹, and a score is then assigned to each private by considering the sum of weighted genetic furnishings previously associated with the phenotype. Studies applying PRS to clinically diagnosed Advert patients have shown a predictive accuracy higher than 80%^12,13, which suggests at that place is potential for PRS to be used as a future clinically valuable tool. PRS have also been utilized to prioritize individuals for screening of rare variants by identifying those with mutual diseases just depression PRS¹⁴.

Here, we apply a PRS derived from a recent, large GWAS in Advertizing⁷ to the U.k. Biobank (UKBB). We perform genetic clan of common variants using individuals belonging to PRS extremes. We analyze these genetic associations alongside the extensive phenotypic and clinical information available in the UKBB.

Results

Polygenic risk scores

Nosotros computed polygenic hazard scores (PRS) for all unrelated genetically divers Caucasian individuals in the UKBB (n = 377,834), based on the summary statistics of a recent AD GWAS⁷. PRS were determined based on 176,316 variants remaining afterwards clumping (Fig. 1).

GWAS: genome-broad association study using Advertising PRS extremes

Using the individuals falling in each PRS extreme (Lower Quantile vs. Upper Quantile in Fig. 1), we performed a GWAS comparison these 2 groups. There was an inflation in the genomic inflation factor [λ = 2.442; meet Supplementary Fig. 1 for a quantile–quantile (QQ) plot], which was expected given the approach of separating individuals based on their genetic adventure.

We identified a total of 246 loci (473 lead SNPs) that met the genome-wide significance criteria (p < 5 × ten^–8) (Supplementary Table 1). We refer to loci by the nearest gene where each lead SNP was annotated, as defined by FUMA.

We identified 23 loci beneath a p value = 1 × 10^–15 that were non present in the base GWAS used for the PRS adding. These are reported in Tabular array i and highlighted in blackness in Fig. 2. Some of these new signals were mapped to genes not previously associated with AD, e.g.: HMGB1P45, RAB23, DIRAS2, SCAPER and TRIM48.

Table ane Top SNPs with p values beneath 1E−15 in UKBB GWAS of Advertizement PRS extremes and non associated with Advertizement in the base of operations GWAS.

Total size table

Variants with p value ≤ 1 × 10^–15 in the UKBB GWAS of PRS extremes non reported as associated in the base of operations GWAS⁷. SNP positions are in GRCh37/hg19. Genomic locus is the index of the genomic chance loci divers by contained atomic number 82 SNPs and maximum distance betwixt their LD block (> 250 kb apart), defined according to FUMA.

We also evidence a comparing betwixt these findings and the main findings reported in the base of operations study⁷ in Table ii. Several meaning results spanned genes that at the time of the base GWAS study, had only been implicated in Advertizement in studies other than typical GWAS: CHRNA6 ¹⁷, ATP7B ¹⁸, IL34 ^nineteen, VAC14 ^xx, KANSL1 ²¹, NOVA2 ²² (Fig. three).

Table 2 Meaning variants reported by Ref.⁷.

Full size table

Of the 44 loci currently associated with Advertizement and nowadays in the GWAS Catalog (version e96_r2019-09-24) we direct replicate 28 (Supplementary Tables ii and 3). The sixteen loci not replicated in our study included half dozen that were not assayed due to low MAF or genotyping call rate of the reported SNPs (HESX1, MEF2C, SHARPIN, SORL1, SUZ12P1/DSG2, ALPK2). The ten loci that remained non-replicated included 3 that were borderline significant in our GWAS (CLNK/HS3ST1, BZRAP1-AS1, and CASS4) and 7 that were non close to genome-wide significance (LCORL, ANKRD31, NME8, NDUFAF6:TP53INP1, PLCG2, CD33, and APP). LCORL, ANKRD31, NDUFAF6:TP53INP1 have only been seen in 1 GWAS each and exercise not accept back up from the most recent GWAS either²³, most likely representing false positives. PLCG2 is a well established locus that is not existence replicated here. APP has been initially shown to be associated past GWAS by Ref.²⁴ and reaches a significance level of p = 1.0 × 10^–12 in Ref.⁵. CD33 has been establish to be pregnant and non-significant past several GWAS studies. The two well-nigh recent GWAS reflect this pattern with the locus showing a significance level of two.21 × 10⁻¹⁰ in Ref.²³ and not showing upwardly in Ref.^v either every bit an established or new locus.

When comparing results with the base GWAS⁷, the well-nigh significant SNPs for ADAMTS4, CR1, HLA-DRB1, CD2AP, ZCWPW1/PILRA, EPHA1, MS4A6A, PICALM, ADAM10, KAT8, SCIMP, and ABCA7 were all more pregnant in this study. Similarly, to the comparing with the loci present in the GWAS catalog, we could not replicate some of the initial findings due to the SNP frequency being lower than our inclusion threshold. Private inspection of these variants revealed several of them had a higher frequency in the loftier PRS grouping than the low PRS, showing the same direction of upshot (Table ii). For example, rs187370608, in TREM2, had a frequency twice as high in the loftier PRS group compared to the low PRS group (MAF: 4.nine × ten^–iii vs. ii.two × 10^–three). Exceptions were rs11218343 in SORL1 and the SUZ12P1 locus that did not bear witness significant differences betwixt groups. In add-on to SNPs that were below our MAF threshold, there were others that we did not replicate, and these were either borderline meaning in our data, or were not farther replicated by more recent AD GWAS. Conversely, some loci that were sub-significant in the base GWAS reached significance in this assay. Some loci, such as IL34, that were not pregnant in the Jansen GWAS, have surpassed the significance threshold in our study and take besides been independently shown to exist associated with Advertising⁵.

Phenotype-based gene set enrichment

To make up one's mind if there were sets of genes associated with other phenotypes enriched in Advert PRS extremes, nosotros performed a gene prepare enrichment assay using FUMA (Supplementary Table 4). In Fig. 4 we report the meridian ten nigh significant GWAS Catalog traits where genes overlap betwixt the GWAS results for each trait and the GWAS results from the AD PRS extremes. To consider the strong issue of the APOE locus we separated results according to the presence (Fig. 4B) or absenteeism (Fig. 4C) of genes located in this locus in the resulting overlapping gene sets.

Genetic correlation

We performed a genetic correlation to determine the relationship to other traits associated with these loci. In Fig. 5 we report the most significant correlated traits when analyzing all datasets available within the "ieu-a" batch available in the OpenGWAS project from the MRC Integrative Epidemiology Unit (IEU) (Fig. 5A). Again, to account for the strong effect of APOE nosotros also conducted this analysis excluding the APOE locus (Fig. 5B). Results for all correlations performed are available in Supplementary Table half dozen.

PheWAS

To decide if whatever of the phenotypes reported in the UKBB dataset were associated with farthermost genetic run a risk for Advertisement and potentially find traits that could be prodromal in Advert, we performed a PheWAS using 1424 traits. Nosotros too performed the same assay to understand if the associations were being driven past APOE, by excluding the APOE locus from the underlying GWAS.

We focused on associations that surpassed the adjusted p value threshold and had a β ≥ |0.v| (Fig. 6, Tabular array 3).

Table 3 PheWAS results for Ad PRS extremes.

Full size table

Family history of AD or dementia (represented past parents and by siblings) was significantly associated with the Advertisement PRS extremes. Interestingly, a proxy to longevity in the parents (female parent and male parent still alive) was also associated with AD PRS extremes only in the opposite direction. Other traits also significantly associated with AD PRS extremes included trend to autumn, fecal incontinence, Parkinson'southward disease in mother, and usage of Gingko forte as medication.

Discussion

Given the difficulty in assembling always larger cohorts of well characterized AD cases and the fact that other genetic loci contributing to disease adventure are still to exist identified, it is important to notice new belittling methods to fully narrate the genetic architecture of Alzheimer's disease. Here we used an alternative arroyo to the typical GWAS past performing an association report on genetic risk extremes for Advertizing. The rationale for this approach is based on the hypothesis that by taking a population of individuals and enriching those that acquit variants that are of take a chance for a phenotype, one is also enriching for other variants associated with that same phenotype. Thus, we are using the farthermost polygenic risk of Ad every bit a surrogate for disease status and the variants identified here are non necessarily associated with Advert itself only with the polygenic risk of Advertizement. In fact, of the 18,892 individuals included in the high PRS extreme, only 365 were reported to have a diagnosis of Ad. Information technology is important to note that we are not using the PRS to predict Advert status; this would suffer from overfitting since the UKBB samples were used in the study in our base of operations summary statistics. In short, we are separating individuals in the UKBB population solely past their PRS for Advertisement, selecting the extremes of this distribution, and then testing their genetic and phenotypic differences.

In this approach, we used 38,000 individuals selected from the PRS extremes obtained for nigh 400,000 individuals in the UKBB. By comparing these genetic adventure extremes, nosotros were able to identify 23 loci with p values of association below one × 10^–fifteen that were not shown as associated with Ad in the base GWAS study. Amidst these, the identification of genome-wide meaning signals at the IL34, ACE and KANSL1 loci—loci that were not meaning in the base of operations GWAS but were subsequently identified to exist associated with AD in independent studies—shows the validity of this arroyo. These results too offer the possibility of auditing the many loci now associated with Advert gamble. Comparing the results from the recent GWAS studying over i million individuals²³ and the previous results by the same group (base of operations GWAS), using largely data on the same samples information technology is interesting to note that 8 loci that were significant in the base GWAS have not been found to be associated with Advert in the larger, more recent study: ADAMTS4, HESX1, HS3ST1, CNTNAP2, KAT8, SUZ12P1 (DSG2), ALPK2 and AC074212.3. Four of these loci are significant in this GWAS of AD PRS extremes: ADAMTS4, CNTNAP2, KAT8 and AC074212.iii. ADAMTS4 is an interesting locus as it has not been associated with Advertisement by whatever other GWAS but shows a very meaning association in this GWAS of AD PRS extremes (p = 6.9 × 10^–18) and functionally is very relevant for the beta-amyloid pathway^25,26. On the other hand, three of the iv loci that were not replicated in this written report (HESX1, SUZ12P1 and ALPK2) all accept top SNPs that were not included in this study due to depression MAF. Nevertheless, at that place is no hint of clan in whatsoever of these loci and they have not been associated with disease in more recent GWAS, indicating these may be imitation positives. Likewise interesting to note is the CD33 locus that has repeatedly been establish to be significantly associated (including in the base of operations GWAS), or to not associate with AD risk in the main GWAS for the disease. These inconsistent results may reverberate an association that is stronger, or only nowadays in some populations, merely tin can also represent a false positive. The most significant variant in this locus in this GWAS of AD PRS extremes reached a p value of 10^–v, supporting the latter. Still, the several studies showing dissimilar furnishings of CD33 in Advertizement, such as an elevated expression in AD encephalon associated with amyloid pathology, affliction progression, and microglial activation may reflect the important part of CD33 biological pathway in AD, nigh likely dependent on TREM2²⁷.

Given the design of this written report, it is not possible to perform a formal replication stage to confirm the novel loci identified that potentially acquaintance with AD chance. Still, some of the loci take now been identified in other GWAS (e.g. WDR12 and DOC2A)^five and many of the genes nominated in the loci take features suggesting a potential role in Advertizement. This recent GWAS also identified GRN and TMEM106B as novel loci for Advert and suggested a continuum between AD and FTD. Interestingly, our results identified several loci that have been previously associated with the risk of Parkinson'southward disease (e.g.: LRRK2, ITKB, CCDC62), simply no loci overlapping with frontotemporal dementia in addition to the MAPT locus. This indicates that instead of a continuum between Advert and FTD at the GWAS risk level, the identification of FTD loci most likely reflected the misclassification in the diagnoses of clinical Advertizement and proxy-AD. Misdiagnoses have e'er been part of GWAS, and these should be more apparent every bit the sample sizes increase with the inclusion of not-and so-well characterized samples.

Both genetic correlation and cistron prepare enrichment analyses identified interesting overlaps. Particularly, gene prepare overlap and genetic correlation can be observed with psychiatric traits such equally schizophrenia, neuroticism, and depression. A previous association study of the shared genetics of autism spectrum disorder, attention arrears-hyperactivity disorder, bipolar disorder, major depressive disorder, and schizophrenia past the Psychiatric Genomics Consortium was ane of the most significant correlations in this analysis. Etiologically and clinically, these psychiatric traits and AD are different diseases. Yet, in many cases, they accept similarities in the patterns of regional brain and biochemical dysfunctions, also as in symptomatology²⁸. Psychotic events are experienced by upward to fifty% of Advertizement patients over the form of their disease²⁹ and, when compared with the general population, individuals with schizophrenia accept a significantly college risk (2–4 times) of developing Advertisement and other dementias^xxx. It is interesting to note that the base GWAS reported a nominally significant genetic correlation between schizophrenia and Advertizing⁷. More recently, past applying a schizophrenia PRS to AD with psychosis, it was shown that psychosis in AD shares some genetic liability with schizophrenia³¹.

It is also interesting that this approach using Advertizement PRS extremes identified enrichment and correlation of genes overlapping with phenotypes such every bit sleep duration and neuroticism. A robust clan of sleep elapsing in middle and old age with the incidence of dementia has recently been established, using the 25-twelvemonth follow-upwards Whitehall II study³². It should likewise be noted that slumber duration is anticorrelated with hazard of Ad in our study. Similarly, neuroticism has been associated with the chance of Advertizement just besides with disease pathology and progression both in sporadic and autosomal dominant disease^33,34,35.

When examining the presence of APOE in these gene sets, the most enriched phenotypes with prior association to APOE included body mass index, AD CSF biomarkers and HDL/LDL levels, and several of these enrichments are further corroborated with evidence for correlation of anthropometric traits in the genetic correlation analyses. This is indicative of the strong issue that APOE has on these phenotypes, but also that the approach to separate individuals based on their AD PRS captures an enrichment of genes directly associated with AD, but too AD-related phenotypes, such as CSF Abeta and tau levels. Excluding the APOE locus from the overlapping genes also showed an enrichment of AD-related phenotypes, but besides of other diseases, such as sarcoidosis and Parkinson'due south disease.

To explore the phenotypes associated with each quantile and potentially find new phenotypes and traits that could be seen as either comorbidity or predicting factors for Advertisement, we also performed a PheWAS, using more than 1000 traits available in the UKBB dataset.

As expected, Ad diagnosis and AD in the family were the most meaning phenotypes associated with the Advertizing PRS extremes comparison. Tendency to fall has been previously shown to be significantly college in a modest cohort of 140 AD patients versus 137 controls³⁶, a result that we replicate in this study. The use of Ginkgo Forte was also significantly associated in these results, which could be a result driven by individuals with a family history of dementia searching for pharmaceutical options to improve or maintain memory. Parental longevity is inversely associated with AD PRS since individuals in the high PRS grouping seemed to have higher bloodshed in their parents. Previous studies accept also reported that individuals with parents who live longer tend to accept a more preserved encephalon construction and lower evidence of AD^37,38.

At that place are a few features of the approach taken in this study that demand to be kept in mind when interpreting these results: as previously mentioned, an extreme PRS for Advertizement does not equate to a clinical diagnosis of AD. The associations described hither are not with Advert itself but rather with the genetic risk for Advertizing. Related to this, high risk individuals may never develop Ad, but they are still genetically predisposed to information technology. It is not possible to easily replicate the results obtained here, given the absence of a similar, independent dataset. Like nigh GWAS, this report also focuses on individuals of European ancestry—a feature of our method that utilizes the largest bachelor "genetically homogenous" publicly available dataset, but an important attribute that is necessary to address in futurity studies³⁹.

Using publicly available data from a previous GWAS on Alzheimer's Illness^vii we computed polygenic hazard scores for all genetically unrelated Caucasians in the UK Biobank cohort. To our knowledge, this is the first report using an Advert PRS to separate individuals purely based on genetic risk, agnostic to disease status. We identified the ii extremes of AD chance from the polygenic gamble distribution and analyzed genetic and phenotypic differences betwixt these groups.

The ability of this unique arroyo allowed us to identify novel associations, non only at loci that were sub-significant in the base study but also at loci that were not suggestively significant. Some of the loci identified here take been recently and independently associated with AD by typical GWAS, validating this approach. Our findings indicate the urgent need of a systematic and comprehensive audit of all loci currently associated with Advertisement risk. The inclusion of loosely characterized samples and the utilise of the same samples and/or data by different GWAS contributes to the difficulty in assessing true loci for the affliction.

In summary, this is the first fourth dimension PRS are used as the only defining characteristic to differentiate groups of individuals to identify novel loci associated with the underlying phenotype. Furthermore, we used phenotype analyses to identify comorbidities, traits, and diseases that can point towards new prodromal characteristics of high genetic take chances for AD.

Methods

Dataset

We used the UKBB cohort, containing 487,409 whole-genome genotyped individuals (version 3)⁴⁰, with about 200,000 of which as well whole-exome sequenced (released in October 2020)⁴¹. This work was conducted equally part of Uk Biobank application number 11036 and follows all applicable guidelines and regulations. Individuals are from the United Kingdom and aged between 40 and 69 at recruitment^forty. We included individuals identified in the UKBB documentation equally genetically defined "Caucasian" and removed individuals with greater than 3rd-caste relatedness to any other sample in the dataset, by applying a Male monarch cutoff of 0.0884 as implemented in the ukbtools bundle (v0.eleven.3).

Polygenic risk score

To derive polygenic risk scores, we applied PRSice-ii⁴² to the summary statistics of 1 of the recent GWAS for AD^seven. Variants with p values below 0.05 in the Advertisement GWAS were selected from the UKBB dataset and filtered to keep only variants with a Hardy–Weinberg equilibrium verbal test p value above 1 × 10^–15, missing phone call rates less than i% and a pocket-size allele frequency of at least 0.one% in the UKBB dataset. We used the following covariates throughout the analysis: sex, year of birth, Townsend deprivation alphabetize at recruitment, genotype measurement batch, and the kickoff x primary components provided by the UKBB. We divers quantiles from the PRS distribution with individuals in the 5% lowest PRS (xviii,892 individuals; 53.8% females, 46.two% males) and the v% highest PRS (eighteen,892 individuals; 54.6% females, 45.4% males). Individuals in the upper quantile will be referred to as "high PRS" individuals, while individuals in the lower quantile will be referred to as "depression PRS" individuals. Additionally, to determine how much of the PRS was dependent on the APOE locus, we calculated PRSs using the complete set of markers and excluding SNPs within one Mb of the nearly significant variant in this locus, while using APOE genotype as a covariate.

Genome-broad association study

We determine which individuals fall in the highest and lowest 5% of the PRS distribution and perform a GWAS using these PRS extremes as the nomenclature of groups in an doubter approach to the clinically defined phenotype. Nosotros adjusted this analysis with the same covariates used in the PRS analysis described higher up. We filtered out all variants with a minimum allele frequency below 5%, Hardy–Weinberg equilibrium exact examination p value below 1 × 10^–6, and missing call rates above i%. Clan analyses were performed using the logistic regression role in PLINK1.9⁴³. We then used FUMA package v1.iii.6a^fifteen to annotate, analyze and interpret the results using the SNP2GENE function. All SNPs prioritized every bit the pb had a p value of less than or equal to 5 × 10^–8.

Additionally, candidate SNPs were included in the notation if they had a maximum p value of 0.05. Meaning SNPs were considered as independent if they had a clumping Rⁱⁱ threshold of at least 0.6 while lead SNPs were prioritized from independent SNPs and only considered every bit such if they had an Rⁱⁱ threshold for the 2d clumping stride of at least 0.one (or if information technology was the same as the first clumping). Nosotros used Stage three of 1000 Genomes (European samples only) as a reference panel to assess linkage disequilibrium.

Genomic aggrandizement was calculated for lambda (λ) in the QCEWAS package in R.

Phenotype-based gene set enrichment

Using results from the GWAS applied to individuals in the farthermost PRS, nosotros performed gene prepare enrichment analyses through GENE2FUNC in the FUMA package v1.3.6a¹⁵. Positional gene mapping aligned meaning SNPs (p value < five × 10⁻⁸) past their location within or immediately upwardly/downstream [± 10 kilobases (kb)] of known cistron boundaries. We report gene sets that had an overlap of at to the lowest degree two genes between the input list of genes (from SNP2GENE) and the gene sets that were significantly enriched at a maximum adapted p value threshold of 0.05. Multiple test correction for gene-set enrichment was performed using the Benjamini–Hochberg (FDR) method⁴⁴.

Genetic correlation

A genetic correlation analysis was performed using LD score regression⁴⁵. We analyzed traits available through the OpenGWAS platform⁴⁶, specifically using the ieu-a batch, which has been well described elsewhere⁴⁷. These summary statistics were filtered to but include datasets with more than 2000 male and female samples, and just those reported in European ancestry groups, yielding 149 datasets. These correlations were also performed in the absence of the APOE locus. All SNPs within the region of 19:45236729–45618959 (hg19) were excluded in this analysis.

Phenome-wide association analysis (PheWAS)

Nosotros used PHESANT—PHEnome Scan Analysis Tool⁴⁸ to perform an automated phenome scan in the UKBB, using the PRS extremes GWAS. This assay was performed including and excluding the APOE locus in the GWAS. Phenotypes with more than 20% missing answers were filtered out. We adjusted for sex, age at recruitment, Townsend deprivation alphabetize at recruitment, genotype measurement batch, and the kickoff ten master components. In addition, nosotros considered phenotype categories with a minimum size of 200 answers and converted fields with multiple instances to categorical (multiple) fields every bit implemented in PHESANT. In total, 1424 traits were analyzed. p values were adjusted for multiple testing correction using Bonferroni.

References

Chouraki, 5. & Seshadri, S. Genetics of Alzheimer's disease. Adv. Genet. 87, 245–294 (2014).

CAS PubMed Article Google Scholar
Gatz, M. et al. Part of genes and environments for explaining Alzheimer illness. Curvation. Gen. Psychiatry 63, 168–174 (2006).

PubMed Article Google Scholar
Nazarian, A. & Kulminski, A. M. Evaluation of the genetic variance of Alzheimer's disease explained past the disease-associated chromosomal regions. J. Alzheimers. Dis. 70, 907–915 (2019).

CAS PubMed PubMed Central Article Google Scholar
Schwartzentruber, J. et al. Author Correction: Genome-wide meta-analysis, fine-mapping and integrative prioritization implicate new Alzheimer's disease risk genes. Nat. Genet. 53, 585–586 (2021).

CAS PubMed Article Google Scholar
Bellenguez, C. et al. New insights on the genetic etiology of Alzheimer's and related dementia. medRxiv (2020).
Wightman, D. P. et al. Largest GWAS (North=1,126,563) of Alzheimer'due south illness implicates microglia and immune cells. bioRxiv https://doi.org/ten.1101/2020.eleven.twenty.20235275 (2020).

Commodity Google Scholar
Jansen, I. East. et al. Genome-wide meta-assay identifies new loci and functional pathways influencing Alzheimer's affliction risk. Nat. Genet. 51, 404–413 (2019).

CAS PubMed PubMed Central Article Google Scholar
Kunkle, B. W. et al. Genetic meta-assay of diagnosed Alzheimer'due south disease identifies new chance loci and implicates Aβ, tau, immunity and lipid processing. Nat. Genet. 51, 414–430 (2019).

CAS PubMed PubMed Primal Article Google Scholar
Perrone, F., Cacace, R., van der Zee, J. & Van Broeckhoven, C. Emerging genetic complexity and rare genetic variants in neurodegenerative encephalon diseases. Genome Med. 13, 59 (2021).

CAS PubMed PubMed Central Commodity Google Scholar
Torkamani, A., Wineinger, N. E. & Topol, E. J. The personal and clinical utility of polygenic take a chance scores. Nat. Rev. Genet. 19, 581–590 (2018).

CAS PubMed Article Google Scholar
Lewis, C. M. & Vassos, E. Polygenic risk scores: From inquiry tools to clinical instruments. Genome Med. 12, 44 (2020).

PubMed PubMed Central Commodity Google Scholar
Escott-Price, V., Myers, A., Huentelman, Grand., Shoai, M. & Hardy, J. Polygenic risk score analysis of Alzheimer's illness in cases without APOE4 or APOE2 Alleles. J. Prev. Alzheimers Dis. half dozen, 16–19 (2019).

CAS PubMed Google Scholar
Chaudhury, South. et al. Alzheimer's disease polygenic take a chance score equally a predictor of conversion from balmy-cognitive damage. Transl. Psychiatry 9, 1–7 (2019).

CAS Article Google Scholar
Lu, T. et al. Individuals with common diseases but with a low polygenic chance score could be prioritized for rare variant screening. Genet. Med. https://doi.org/10.1038/s41436-020-01007-7 (2020).

Commodity PubMed Google Scholar
Watanabe, K., Taskesen, E., van Bochoven, A. & Posthuma, D. Functional mapping and note of genetic associations with FUMA. Nat. Commun. eight, 1826 (2017).

ADS PubMed PubMed Cardinal Article CAS Google Scholar
Pruim, R. J. et al. LocusZoom: Regional visualization of genome-wide association scan results. Bioinformatics 26, 2336–2337 (2010).

CAS PubMed PubMed Cardinal Article Google Scholar
Altimiras, F. et al. Brain transcriptome sequencing of a natural model of Alzheimer's illness. Front end. Aging Neurosci. 9, 64 (2017).

PubMed PubMed Fundamental Article CAS Google Scholar
Squitti, R., Siotto, M., Arciello, G. & Rossi, L. Not-ceruloplasmin leap copper and ATP7B factor variants in Alzheimer'south disease. Metallomics 8, 863–873 (2016).

CAS PubMed Article Google Scholar
Walker, D. G., Tang, T. M. & Lue, 50.-F. Studies on colony stimulating factor receptor-1 and ligands colony stimulating factor-1 and interleukin-34 in Alzheimer's illness brains and human being microglia. Front. Aging Neurosci. 9, 244 (2017).

PubMed PubMed Key Article CAS Google Scholar
Drange, O. Thousand. et al. Genetic overlap between Alzheimer'south disease and bipolar disorder implicates the MARK2 and VAC14 genes. Front. Neurosci. 13, 220 (2019).

PubMed PubMed Fundamental Commodity Google Scholar
Logue, M. W. et al. Targeted sequencing of Alzheimer illness genes in African Americans implicates novel risk variants. Front end. Neurosci. 12, 592 (2018).

PubMed PubMed Central Commodity Google Scholar
Tollervey, J. R. et al. Assay of alternative splicing associated with crumbling and neurodegeneration in the man brain. Genome Res. 21, 1572–1582 (2011).

CAS PubMed PubMed Central Article Google Scholar
Wightman, D. P. et al. A genome-wide association written report with ane,126,563 individuals identifies new risk loci for Alzheimer's affliction. Nat. Genet. 53, 1276–1282 (2021).

CAS PubMed Article Google Scholar
Moreno-Grau, Southward. et al. Genome-broad clan analysis of dementia and its clinical endophenotypes reveal novel loci associated with Alzheimer's disease and 3 causality networks: The GR@ACE project. Alzheimers. Dement. 15, 1333–1347 (2019).

PubMed Article Google Scholar
Satoh, Yard., Suzuki, N. & Yokota, H. ADAMTS-4 (a disintegrin and metalloproteinase with thrombospondin motifs) is transcriptionally induced in beta-amyloid treated rat astrocytes. Neurosci. Lett. 289, 177–180 (2000).

CAS PubMed Article Google Scholar
Tomita, T. et al. Identification of ADAMTS4 as an APP-cleaving enzyme at 669 site in the APP669-711 product pathway. Alzheimers. Dement. 16, e039194 (2020).

Google Scholar
Griciuc, A. et al. TREM2 acts downstream of CD33 in modulating microglial pathology in Alzheimer's illness. Neuron 103, 820-835.e7 (2019).

CAS PubMed PubMed Fundamental Article Google Scholar
White, Thousand. E. & Cummings, J. L. Schizophrenia and Alzheimer's disease: Clinical and pathophysiologic analogies. Compr. Psychiatry 37, 188–195 (1996).

CAS PubMed Article Google Scholar
Ropacki, S. A. & Jeste, D. V. Epidemiology of and take chances factors for psychosis of Alzheimer's disease: A review of 55 studies published from 1990 to 2003. Am. J. Psychiatry 162, 2022–2030 (2005).

PubMed Article Google Scholar
Ribe, A. R. et al. Long-term risk of dementia in persons with schizophrenia: A Danish population-based cohort written report. JAMA Psychiat. 72, 1095–1101 (2015).

Article Google Scholar
Creese, B. et al. Examining the association betwixt genetic liability for schizophrenia and psychotic symptoms in Alzheimer's illness. Transl. Psychiatry nine, 273 (2019).

PubMed PubMed Cardinal Article Google Scholar
Sabia, S. et al. Association of sleep duration in heart and old age with incidence of dementia. Nat. Commun. 12, 2289 (2021).

ADS CAS PubMed PubMed Central Commodity Google Scholar
Schultz, Southward. A. et al. Association between personality and tau-PET binding in cognitively normal older adults. Encephalon Imaging Behav. 14, 2122–2131 (2020).

PubMed PubMed Fundamental Article Google Scholar
Duberstein, P. R. et al. Personality and take chances for Alzheimer's disease in adults 72 years of age and older: A six-year follow-up. Psychol. Aging 26, 351–362 (2011).

PubMed PubMed Central Article Google Scholar
Aschenbrenner, A. J. et al. Relationships betwixt big-five personality factors and Alzheimer'due south disease pathology in autosomal dominant Alzheimer's disease. Alzheimers. Dement. 12, e12038 (2020).

Google Scholar
Dev, K. et al. Prevalence of falls and fractures in Alzheimer's patients compared to general population. Cureus xiii, e12923 (2021).

PubMed PubMed Primal Google Scholar
Murabito, J. Thousand. et al. Parental longevity is associated with knowledge and brain ageing in middle-aged offspring. Age Ageing 43, 358–363 (2014).

PubMed Article Google Scholar
Lipton, R. B. et al. Infrequent parental longevity associated with lower hazard of Alzheimer's disease and memory turn down. J. Am. Geriatr. Soc. 58, 1043–1049 (2010).

PubMed PubMed Central Commodity Google Scholar
Dehghani, N., Bras, J. & Guerreiro, R. How understudied populations have contributed to our agreement of Alzheimer'south illness genetics. Brain https://doi.org/10.1093/encephalon/awab028 (2021).

Article PubMed PubMed Central Google Scholar
Bycroft, C. et al. The UK Biobank resource with deep phenotyping and genomic data. Nature 562, 203–209 (2018).

ADS CAS PubMed PubMed Fundamental Article Google Scholar
Szustakowski, J. D. et al. Advancing man genetics inquiry and drug discovery through exome sequencing of the Britain Biobank. bioRxiv https://doi.org/ten.1101/2020.11.02.20222232 (2020).

Commodity Google Scholar
Choi, S. W. & O'Reilly, P. F. PRSice-2: Polygenic risk score software for biobank-calibration information. Gigascience eight, giz082 (2019).

PubMed PubMed Central Article Google Scholar
Chang, C. C. et al. 2nd-generation PLINK: Ascension to the claiming of larger and richer datasets. Gigascience 4, 7 (2015).

PubMed PubMed Central Article CAS Google Scholar
Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc. 57, 289–300 (1995).

MathSciNet MATH Google Scholar
Bulik-Sullivan, B. et al. An atlas of genetic correlations beyond human diseases and traits. Nat. Genet. 47, 1236–1241 (2015).

CAS PubMed PubMed Key Article Google Scholar
Elsworth, B. et al. The MRC IEU OpenGWAS data infrastructure. bioRxiv https://doi.org/10.1101/2020.08.10.244293 (2020).

Article Google Scholar
Hemani, Thousand. et al. The MR-Base platform supports systematic causal inference across the human phenome. Elife 7, e34408 (2018).

PubMed PubMed Fundamental Commodity Google Scholar
Millard, L. A. C., Davies, Northward. M., Gaunt, T. R., Smith, Yard. D. & Tilling, K. PHESANT: A tool for performing automated phenome scans in UK Biobank. Cold Spring Harb. Lab. https://doi.org/10.1101/111500 (2017).

Commodity Google Scholar

Download references

Acknowledgements

This research has been conducted using data from Britain Biobank, a major biomedical database (www.ukbiobank.air conditioning.uk) nether application number 11036. The authors are grateful for funding support from the National Constitute on Crumbling of the National Institutes of Health nether Honour Number R01AG067426 and the Van Andel Inquiry Institute.

Writer information

Affiliations

Department of Neurodegenerative Scientific discipline, Van Andel Research Institute, 333 Bostwick Ave. N.E., Grand Rapids, MI, 49503-2518, USA

Catarina Gouveia, Elizabeth Gibbons, Nadia Dehghani, James Eapen, Rita Guerreiro & Jose Bras
Division of Psychiatry and Behavioral Medicine, Michigan Country University College of Human Medicine, Grand Rapids, MI, United states of america

Rita Guerreiro & Jose Bras

Contributions

J.B. and R.G. conceived the thought, supervised the work, and drafted the manuscript. C.G. performed the assay and drafted the manuscript. E.G., N.D. and J.E. performed analysis and interpretation of results. All authors discussed the results and contributed to the final manuscript.

Corresponding writer

Correspondence to Jose Bras.

Ethics declarations

Competing interests

The authors declare no competing interests.

Boosted data

Publisher's note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Rights and permissions

Open Access This article is licensed nether a Creative Commons Attribution iv.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give advisable credit to the original author(due south) and the source, provide a link to the Creative Commons licence, and bespeak if changes were made. The images or other third party material in this article are included in the article'due south Artistic Commons licence, unless indicated otherwise in a credit line to the fabric. If material is non included in the article's Creative Commons licence and your intended apply is not permitted by statutory regulation or exceeds the permitted use, yous will need to obtain permission direct from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/four.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gouveia, C., Gibbons, E., Dehghani, N. et al. Genome-broad association of polygenic gamble extremes for Alzheimer's disease in the UK Biobank. Sci Rep 12, 8404 (2022). https://doi.org/10.1038/s41598-022-12391-2

Download citation

Received: 03 December 2021
Accepted: 10 May 2022
Published: 19 May 2022
DOI : https://doi.org/10.1038/s41598-022-12391-two

Comments

By submitting a comment you concord to abide by our Terms and Community Guidelines. If you lot find something abusive or that does not comply with our terms or guidelines please flag it equally inappropriate.