There are many active research projects accessing and applying shared ADNI data. Use the search above to find specific research focuses on the active ADNI investigations. This information is requested annually as a requirement for data access.
| Principal Investigator | |
| Principal Investigator's Name: | Jongsun Jung |
| Institution: | Syntekabio |
| Department: | Genome Data Integration Center |
| Country: | |
| Proposed Analysis: | Research Proposal We write this mail to inquire possible access to the whole exome sequencing and neuroimaging data for 800 AD cohorts. We have developed a deep learning pipeline specialized in human genome big data analyses. The deep learning pipeline includes standard tools such as NGStoolkit, BWA-MEM and GATK as well as in-house packages to calculate SNVs, CNVs, multi-genome integration and tools for deep learning (Convolutional Neural Network)-based Patient stratification and Biomarker discovery. In particular, our deep learning approach is performed by converting one dimensional nucleotide sequences to filtered 2D images of Hilbert curve (Hilbert is one of great mathematicians in 1890), reported by Vivien Marx, Visualizing epigenomic data, Nature Methods 12, 2015: 499–502; Simon Anders, Visualization of genomic data with Hilbert curve, Bioinformatics, 2009. Its performance with our deep learning test result appears par to other publically available tools in biomarker discovery (in writing). In addition, the CNV pipeline for the deep learning setting is also more powerful than existing tools in finding deletions and duplications of exons in the ranges of a few hundred to over a million bases. Using the in-house tools, we have analyzed exome data for >600 cohorts, comprised of 418 patients and 173 healthy controls collected by Parkinson’s Progression Markers Initiatives. Mutation burden was not particularly higher among the patient group compared to the healthy control group. Only 10 of 414 patients had causative mutations that were clearly associated with PD. Copy number variations in multiple exons of critical proteins causing rare genetic diseases were discovered in 157 patients ranging from one to nine. Both presence and numbers of CNVs negatively correlated to age at onset of PD but patients did not show common biomarkers. We suspect that the faulty proteins originated from CNVs might have triggered unknown, diverse underling PD mechanisms. Based on the observation among PD patients, we hypothesized that neurodegenerative diseases, including AD, are associated with dilapidation of lysosomal functions overloaded by faulty proteins generated from the genes with partial deletions or duplications of exons ( in writing) To test the hypothesis, we made a research consortium among several research hospitals including the department of geriatric psychiatry of the Catholic University of the Korea the Ajou University of Korea, the Seoul National University of Korea, the KOREA university of Korea, Yonsei University of Korea, the CHA university of Korea, the Hallym university of Korea, and Syntekabio in S. Korea. We wish to have the whole genome data and neuroimaging data for the AD cohorts to analyze particularly stratified biomarkers with CNVs and SNVs together among AD patients. The consortium plans to apply research grants on AD research at Korea and include the analysis results as preliminary data. For your information, Syntekabio is a research oriented company partially owned by Electronics and Telecommunications Research Institute in S. Korea. The company compiled the genome data produced by Korean Genome project, 1000G, and ICGC. We have supercomputing technologies and can easily handle whole genome sequences of several thousand people at a time. We would appreciate it if you could share the whole genome data for the AD cohorts in the form of either Fastq or BAM files. Please let us know conditions and regulations to access the data |
| Additional Investigators | |
| Investigator's Name: | Yangrae Cho |
| Proposed Analysis: | We downloaded part of ADNI images, patient logs, and genome sequence data. We had analyzed patient logs first to classify patients based on the speed of disease progression from MCI to dementia. One group was composed of patients whose symptom was changed from MCI to dementia within 36 months. The others were stayed in MCI states for over 60 months either progressed or not to MCI. We found that the switch from MCI to dementia occurred within a short time frame after hippocampal region was deteriorated. We plan to analyze all available images to confirm the initial discovery. In addition, we have tried to analyze the genome sequences to discover combination of single nucleotide variants with frequent occurrence among fast progressing patients compared to slowly progression ones. |
