Poster Presentations: Tuesday, July 18, 2017
as covariates for logistic regression combined with PRS. Results: There were a total of 39,146 SNPs common between both datasets. The best threshold for generating scores in the target dataset was P ¼ 0.004, when R2 ¼ 0.015, using information from 506 SNPs. Scores range from -0.008 to +0.032, with means for controls and cases being 0.003 and 0.008 respectively. Sensitivity and specificity were calculated as 0.542 and 0.741 respectively; AUC was 0.686. Analysis of the 506 SNPs highlighted 32 variants in 11 GWAS genes, 3 being GWAS hits, and 4 within the APOƐ region, also implicated in LOAD. Logistic regression produced an AUC of 0.728. Conclusions: The results show that PRS is higher on average in sEOAD cases than controls, although there is still overlap amongst the whole cohort. This suggests that sEOAD is on a spectrum of AD, and that effective variants can be present in controls without leading to AD. The presence of GWAS hits also confirms the role of these in genes in AD.
P3-111
NOVEL CANDIDATE GENES FOR DEMENTIA WITH LEWY BODIES
Tatiana Orme1, Rita Guerreiro1, Isabel Santana2, Jose T. Bras1, 1 UCL Institute of Neurology, London, United Kingdom; 2 Dementia Clinic, Centro Hospitalar e Universitario de Coimbra and Faculty of Medicine, Universidade de Coimbra, Coimbra, Portugal. Contact e-mail:
[email protected] Background: Dementia with Lewy Bodies (DLB) is a neurode-
generative disease that is estimated to affect approximately 5% of people aged over 85. Despite the prevalence of the disease, much of the genetic etiology of DLB is yet to be elucidated. Although most cases are sporadic, there is a genetic component to the disease and so far, only genes previously associated with Alzheimer’s disease (AD) and Parkinson’s disease (PD) have been shown to increase susceptibility for DLB. To date, only a handful of families with DLB have been available for study, and no disease causing mutations have been identified. Here, we performed a comprehensive genetic analysis of 2 cousins clinically diagnosed with DLB with the aim of finding a potential pathogenic mutation responsible for the phenotype. Methods: Whole exome sequencing was conducted in both cases. No other family members were available for study. We focussed our analysis on rare, coding variants that would follow either recessive or a dominant modes of inheritance. Results: No known mutations were found in known Alzheimer’s and Parkinson’s disease genes. We refined the initial list of 80,500 variants with the same genotype present in both cousins, by focussing on rare (AF from 0-8.86E-06 in gnomAD) medium or high impact variants, in genes known to be expressed in the brain. This lead to the identification of 9 candidate variants. In addition, we analysed the frequency of these variants in independent cohorts of DLB, AD and PD exome sequencing datasets. The variants were further assessed using variant effect predictors such as CADD, and burden tests of the genes were performed in the DLB dataset. Conclusions: Familial forms of DLB are exceedingly rare. Here we present a family with two cousins clinically diagnosed with DLB that underwent detailed genetic analyses. We identified 9 variants that are potentially involved in the disease. Making this data available will enable other researchers studying the same phenotype to better analyse their data.
P3-112
P977
INVESTIGATING GENETIC VARIATION IN ALZHEIMER’S DISEASE USING WHOLE-EXOME SEQUENCING
Tulsi Patel1, Keeley J. Brookes1, Tamar Guetta-Baranes1, Sally Chappell1, Rita Guerreiro2, Jose T. Bras2, John Hardy2, Paul T. Francis3, Kevin Morgan1, 1University of Nottingham, Nottingham, United Kingdom; 2UCL Institute of Neurology, London, United Kingdom; 3King’s College London, London, United Kingdom. Contact e-mail:
[email protected] Background: Alzheimer’s
disease (AD) is an incurable neurodegenerative disorder; in which the death of brain cells characteristically result in memory loss and cognitive decline. It is the most common form of dementia, affecting around 850,000 people in the UK. Sporadic late-onset AD (LOAD) accounts for 95% of all cases and is genetically complex in nature. It is believed that combinations of genetic and environmental factors are at play. Recent genome-wide association studies (GWAS) have uncovered over 20 new gene candidates for AD risk, however these exhibit small effect sizes. We are now utilising next generation sequencing (NGS) to explore the contribution made by rare variants (MAF<5%). Recently this approach highlighted the role of TREM2 and SORL1 variants in AD risk and emerging NeuroX chip and NGS data is set to generate more genes of interest. Methods: DNA was extracted from post-mortem brain tissue obtained from the BDR for healthy and diseased individuals. Whole-exome sequencing was performed on 292 samples, including 133 AD cases, 53 controls and 106 other phenotypes. Samples were screened for mutations in APP, PSEN1 and PSEN2 to distinguish early-onset AD cases and known LOAD risk genes. Using a combination of bioinformatics and statistical tools variants were tested for association with AD and their functionality assessed in silico. Results: Presently over 340,000 variants have been identified in the whole-exome dataset, with 18% predicted to be novel. Coding variants account for w55% of all variants, with >50% of these being missense or frameshift mutations. A total of 653 SNPs were found in GWAS nominated genes. APOEε4 SNP rs429358 reached genome-wide significance (p<5x10-8, OR¼6.5) in single SNP association without adjusting for covariates. Conclusions: It is well known that APOEε4 significantly increases AD risk. Although no other SNPs reached genome-wide significance, it is likely that important variants could lie below this threshold. Therefore AD risk gene variants and other neurodegenerative disease risk genes will also be investigated further.
P3-113
MUTATIONAL ANALYSIS OF PRNP IN EARLY ONSET ALZHEIMER’S DISEASE IN KOREA
Vo Van Giau1, Eva Bagyinszky2, Kyu Hwan Shim3, Youngsoon Yang4, Young Chul Youn5, Seong Soo An3, Sang Yun Kim6, 1Gachon University, Seongnam, Republic of South Korea; 2Gachon University, Seongnam, Republic of South Korea; 3Department of Bionano Technology, Gachon University, Seongnam, Republic of South Korea; 4Department of Neurology, Veterans Hospital, Seoul, Republic of South Korea; 5Chung-Ang University Hospital, Seoul, Republic of South Korea; 6Seoul National University Bundang Hospital, Sungnam, Republic of South Korea. Contact e-mail:
[email protected]