NHGRI/DIR Bioinformatics and Scientific Programming Core

点击翻译按钮获取中文摘要

The NHGRI Bioinformatics and Scientific Programming Core actively supports the research being performed by NHGRI/DIR investigators by providing expertise and assistance in bioinformatics and computational analysis. The Core facilitates access to specialized software and hardware, develops generalized software solutions that can address a variety of questions in genomic research, develops database solutions for the efficient archiving and retrieval of experimental and clinical data, disseminates new software and database solutions to the genome community at-large, collaborates with NHGRI researchers on computationally-intensive projects, and provides educational opportunities in bioinformatics to NHGRI Investigators and trainees. The majority of engagements between the Bioinformatics and Scientific Programming Core and DIR investigators are focused on collaborative interactions intended to advance specific research projects. The support provided for these projects includes not only data analysis but also related efforts focused on data collection and dissemination through the public NHGRI/DIR Web site (http://research.nhgri.nih.gov). Responding to the continued demand for variant calling on human genome and exome data, the Core has maintained and updated a GATK-based pipeline that builds upon best practices published by the Broad Institute. This standardized and validated pipeline is currently being used in the context of The Genome Ascertainment Consortium (TGAC) effort; the goals of this effort are to improve our overall understanding of the phenotypic consequences of genetic variation and to predict phenotypes from genotypes. To that end, this pipeline has facilitated the creation of a uniformly processed and formatted genotype callset across multiple cohorts, based on data from multiple sources. We recently processed 550 exome samples from NIAID, which were added to 1,500 exome samples from the ClinSeq cohort and 4,600 genomes from the INOVA Translational Medicine Institute. The pipeline continues to be updated with new software versions and more recent releases of human genome sequences. The pipeline has also been optimized to take advantage of the Biowulf high-performance computing environment, parallelizing the per-sample processing steps and making use of local SSD storage on nodes (as available) to increase speed and reduce network overhead. Going forward, this efficient pipeline will allow for the re-calling of data from this growing cohort of individuals who have agreed to be re-contacted for secondary phenotyping studies, with the increased sample sizes affording greater power to discover important phenotype/genotype associations. Alongside this effort, the Core has developed an interactive browser for visualizing aggregate exome and genome data from the aforementioned TGAC cohorts, using the gnomAD codebase as its foundation. In the past year, we updated the TGAC browser to integrate population allele frequencies from the gnomAD repository derived from 125,000 exomes. We also integrated data for the human X and Y chromosome. The Core has also focused on generating an in-house somatic variant calling pipeline for use in mosaic and cancer somatic variant calling. This pipeline leverages the initial alignment stages of the existing germline pipeline, but in the variant calling step uses Samtools mpileup and Varscan to produce a highly sensitive variant caller capable of detecting alleles present in only 5% of reads in a sample. Key to the utility of this pipeline are detailed quality and read count statistics broken down by strand direction, allowing the scientific end-user to create a custom filtering strategy. This pipeline has already been applied to two projects to date: the detection of mosaicism and the detection of mitochondrial genome heteroplasmy in cell free DNA. Additional projects include the development of computational methods to analyze RNA-seq data obtained from the zebrafish translatome, assessing mRNAs that are being actively translated; implementation of a gene prediction pipeline for the preliminary annotation and analysis of the Hydractinia genome; significant expansion of the Mnemiopsis Genome Project Portal to include in situ images, temporal developmental expression profiles, and single-cell expression data; single-cell analysis of iPSC-derived neural progenitors from individuals with Down syndrome; bioinformatics support for a DDIR Directors Challenge project that is part of the Genomic Ascertainment Cohort (TGAC), including implementation of the gnomAD variant browser, for establishing a shared genomic ascertainment cohort of at least 10,000 individuals whose genomes or exomes have been sequenced and are recallable for secondary phenotyping studies; determining if individuals with Proteus syndrome contain small duplications in exon 4 of AKT; annotation of samples from the TGAC cohorts with HLA genotypes and integration of results into the gnomAD browser; using genotype data, assessing familial relationships among TGAC samples to identify relevant individuals for phenotypic follow-up; re-calling variants from 557 exomes from NIAID and incorporating them into the gnomAD browser; development of a website to return negative secondary findings to participants from the A2 ClinSeq cohort; RNA-seq analysis of whole blood from sickle cell patients to identify differentially expressed genes in individuals with and without leg ulcers; maintenance of a public web browser and BLAT interface for the goldfish (Carassius auratus) genome assembly, based on the UCSC Genome Browser; analysis of single-cell RNA-seq and ATAC-seq data obtained from zebrafish sensory hair cells; as part of the FUSION Project, assessing the feasibility of using singe-cell RNA-seq to interrogate the transcriptomes of pancreatic islet cells obtained from post autopsy tissue, as well as investigating the genetic basis of type 2 diabetes disease risk through the use of single-cell and/or single-nuclei RNA-seq technology, with the goal of interrogating the transcriptomes of the individual cells that comprise the pancreatic islet; updating the Skippy web server to comply with security regulations, as well as including additional complementary tools for splicing prediction; performing isoform expression profiling of pan-cancer datasets in TCGA; performing RNAseq analyses in peripheral blood from patients with mitochondrial disease; continuing maintenance of a customized database and web interface for storing and computing on genomic data from dogs; implementing a GEMINI database with data from 750 dogs to allow complex searches of dog genotypes; designing and implementating of surveys that assess the health of pet dogs whose DNA samples have been submitted to scientific studies; performing RNA-seq analysis of fibroblasts with different cyclodextrin treatments and LysoTracker staining profiles to discover genetic modifiers of Niemann-Pick disease type C (NPC); performing RNA-seq analyses of post-mortem brain tissue to compare neuronal gene expression in youths with a history of ADHD against matched controls in order to establish a neuronal transcriptome and determine the genes and neural gene networks that influence the development of ADHD; performing RNA-seq analysis of the developing fetal brain to understand the effect of GM2 gangliosidosis on gene expression and to determine the regions of the brain most significantly dysregulated by storage of gangliosides; and identifying integration sites of AAV in mouse and human genomes and developing methods to characterize the clustering and locations of the integration sites.