Jumping on Next-Gen Sequence Data Analysis

Best Practices Winner: National Center for Genome Research
Project: NCGR Schizophrenia Genome Project
Category: Basic Research
Nominated By: SAS Institute

By Kevin Davies

July 22, 2009 | In 2007, the National Center for Genome Resources (NCGR) in Santa Fe, New Mexico, best known for the development of bioinformatics tools, established the Schizophrenia Genome Project. Taking Illumina GA sequence data, NCGR scientists used a combination of homegrown Alpheus software and JMP Genomics (from SAS) to develop a streamlined workflow for the acquisition, analysis, and management of huge amounts of next-generation sequencing data—in this case, mRNA.

“This was the first time that these technologies had been integrated for a case-control study,” says Faye Schilkey, associate director of NCGR’s sequencing center. The project offers important insights into a devastating disease and the development of a sophisticated and accessible data analysis pipeline for translational next-gen sequencing projects for groups who might not have the resources of the major genome centers.

Because of the notoriously complicated genetics of schizophrenia, NCGR scientists under director Stephen Kingsmore examined variations in gene expression between cases and controls. The first phase of the project was published last November in PLoS ONE.
“We used mRNA sequencing, which typically hadn’t been done before,” says Schilkey. Using Illumina GA sequencers, the NCGR team generated 16.7 billion bases (473 million reads) of shotgun DNA sequences of cDNA from post-mortem cerebellar cortex of 14 patients and six controls. Those reads were analyzed using Alpheus, which NCGR calls a “web-based cyberinfrastructure platform for pipelining, visualization, and analysis of gigabase-scale medical resequencing studies.”

The sequence reads were aligned to some 33,000 transcripts in each sample, which in turn were used to generate digital gene expression values. Using Alpheus, NCGR identified gene expression differences, splice site differences, and sequence variants (single nucleotide polymorphisms) of more than 33,000 transcripts, while minimizing false positives. (The web portal enables investigators worldwide to explore the results in a highly flexible manner, without the need for a massive local computational infrastructure or advanced bioinformatics expertise.)

Next, NCGR staff adapted Alpheus to export the variant or digital expression data to JMP Genomics statistical software from SAS for visualization and statistical analysis. JMP Genomics is better known for microarray analysis—it produces interactive graphical representations of large volumes of exported data, and is particularly useful for identifying outlier samples, and it worked right out of the box for the analysis of next-gen sequence expression data based on read counts. A bonus was that JMP Genomics could detect and visualize the sensitivity of the Illumina GA. “The dynamic range is much greater than that of arrays. Illumina reads mass units of cDNA and a zero is a true zero compared to arrays where approximately 30% are absent calls due to array hybridization noise,” says Schilkey.

NCGR analysts used principal components analysis and hierarchical clustering to assess the data. The variance attributable to disease status was higher for the Illumina digital expression data than from conventional array analysis. “Visualization tools, such as Principal Component Analysis, readily separated the cases and controls, we spotted differences right away,” says Schilkey. Given more than 11,500 schizophrenia candidate genes in the literature, the Illumina/Alpheus/JMP Genomics pipeline revealed “23 genes with altered expression and involvement in presynaptic vesicular transport, Golgi function, and GABAergic neurotransmission define a unifying molecular hypothesis for dysfunction” in schizophrenia.

The project represents the first large-scale case-control study to utilize mRNA sequencing, yielding novel insights into the molecular mechanisms underpinning this disease. NCGR believes the Alpheus/JMP Genomics pipeline offers a turnkey approach for visualization, results identification, and statistical analysis of next-gen sequencing data. While next-gen sequencing technologies have democratized genome sequencing, Kingsmore points out that “visualization, analysis, and discovery in massive sequence sets remains limited to a few centers nationwide. Alpheus with JMP Genomics has been designed to provide end-to-end democratization of genome, methylome, and transcriptome analysis.”

The Best Practices award will help garner attention from scientists who may not enjoy the same access to resources as a major genome center. The Alpheus/JMP Genomics combination provides a data management, visualization, analysis, and statistical framework for the application of next-generation sequencing data to hypothesis-based, investigator-initiated experiments. Alpheus, which is available as a service, applies vast computational resources with intuitive web-based visualization and query to help relieve the analysis bottleneck, while JMP Genomics delivers robust statistical results. The pairing provides an excellent tool for individual researchers who perhaps lack the extensive bioinformatics expertise required to analyze large-scale sequencing projects.

Schilkey and Kingsmore are delighted to have demonstrated that the workflow works. “More people will catch on, I expect, once they get the volumes of data. We’re a little ahead of the curve,” says Schilkey. “This is a soup-to-nuts solution for next-generation sequence analysis.” Near term additions to the project will be to incorporate Ingenuity’s Pathways Analysis tools and to add functionality to identify splice isoform differences between cases and controls.