Java and Jive Breakout Discussion Groups

Tuesday, August 14 - 8:30 am

Table 1: Sequence Assembly and Analysis - What Does the Future Hold?

Tom Schwei, Vice President & General Manager, DNASTAR

• What is the impact of nanopore technologies?
• De novoassembly of anything and everything – what will it take?
• Metagenomics – a whole new ballgame?
• Curing cancer – what data, tools, and capabilities are needed?

Table 2: Utilization of Genetic Variants Derived from NGS Data for Clinical Research and Applications

Ming Yi, Ph.D., IT Manager, Functional Genomics/Bioinformatics Support Group, Advanced Biomedical Computing Center, SAIC-Frederick, FNLCR

• What tool(s) are you using and like to choose in the future to call variants from NGS data (e.g.,Exome-Seq, Whole Genome Seq)?
• Which one is better: SNPs derived from array platform vs. from NGS platform?
• Is the quality (e.g., error rate) of variants (SNPs, Indels, or SVs) detected from Exome-Seq data good enough for genetic analysis (e.g., linkage analysis)? Any critical limitation?
• What are the Pros and Cons of variants detected from RNA-Seq data vs. from Exome-Seq data? What tool(s) do you use to call variants from RNA-Seq data?
• What are the limitations in the field for the SNP annotation (e.g., dbSNPs, 1KG, HapMap) and SNP impact assessment on proteins (e.g., SIFT, Polyphen)?
• Are the variants derived from NGS data ready for clinics?

Table 3: Variable Human Genomes: How to Definea Unique Reference Genome?

Suganthi Balasubramanian, Ph.D., Associate Research Scientist, Molecular Biophysics and Biochemistry, Yale University

• Is it valid to continue the use of the current human reference genome as the basis for comparison given the avalanche of human genome sequences that are coming?
• Why should we have one reference genome?
• Given the short read lengths, are de novoassemblies a feasible alternate option?
• What are the shortcomings of using one reference genome, especially for clinical Dx and Rx?
• If a reference sequence is essential, what are the ways to improve or derive a new reference human genome?

Table 4: Comparison of High and Low Coverage Sequencing for Variant Detection in Human Genomes

Jeffrey Rosenfeld, Ph.D., Scientific Programmer, University of Medicine and Dentistry of New Jersey

• What are the cost differentials between high and low coverage sequencing?
• What are the benefits and limits of each of the approaches?
• What can we learn from the 1000 Genomes Project about low-coverage sequencing?
• Is exome sequencing a suitable middle-ground between high and low-coverage?

Table 5: Next-Generation Sequencing Data Analysis Protocol for Clinical Applications

May Dongmei Wang, Ph.D., Associate Professor, Biomedical Engineering, Electrical and Computer Engineering, Hematology and Oncology, Winship Cancer Institute, Georgia Institute of Technology

• Is there a best-practice NGS data analysis pipeline that is suitable for anyspecific clinical application?
• How would prevalence affect data analysis classifier for clinical application?
• What is the best performance metric for choosing NGS data classifier?
• What are the top three factors affecting the NGS data analysis outcome?

Table 6: Defining the Transcriptional Landscape through RNA-Seq: from Reads to Isoforms

Dario Motti, Ph.D., Postdoctoral Fellow, The Miami Project to Cure Paralysis, University of Miami, Miller School of Medicine

• Comparing RNA-Seq to other tools: is RNA-Seq the best way to identify and quantify isoform expression?
• How do different isoform-assembly and quantification algorithms compare?
• How do we decide what expression level identifies an isoform as reliably present?
• What is the best way to validate the presence of predicted isoforms?
• Increasing sequencing depth reveals isoforms expressed at very low levels--are these biologically relevant?
• How do we compare among datasets and define an isoform as "novel"?

Table 7: After Sequencing: The Costs of Associating Functions to Genes in Bioinformatic Analyses

Melanie Lehman, Ph.D., Research Fellow, Institute of Health and Biomedical Innovation, Queensland University of Technology

• Is the term ‘gene’ still useful with our current understanding of transcriptome/proteome complexity or is it impeding biological discovery?
• How do overlapping alternative protein-coding and non-coding transcripts alter downstream pathway and functional analyses?
• What tools/databases are you using to match genome variations and alternative transcripts to protein domains to infer biological function?
• How do we include the regulatory roles of RNA (e.g. non-coding transcripts and UTRs) in functional analyses?

Table 8: Accuracy and Validation of Targeted Sequencing Analysis Pipelines

Attila Berces, Ph.D., CEO Omixon Biocomputing

• What are the best ways to validate an NGS analysis pipelines for human targeted sequencing? What are the sources of reference data?
• Target genes differ in the rate of polymorphism, repetitive elements, pseudogenes, GC content, and the typical mutations, and the sequencing runs differ by target amplification, and sequencer. How general and how specific validation should be?
• What artifacts arise in hybridization- and PCR amplification-based assays?
• What false positive/negative rate would be clinically acceptable for some of the common gene targets: BRCA, CFTR?
• What are the advantages and disadvantages of simulation based validation?

Table 9: NGS Data Storage – Best Practices Discussion

Will McGrath, Strategic Marketing and Technology Partnerships, Quantum Corporation

• What approaches have you taken for storing NGS data?
• Do you treat all data equally or have you taken steps to tier or archive older NGS data?
• Do you save all raw data off the instrument or mainly BAM and certain fastq and variant files? 
• Are you currently using or considering the use of the cloud for long term data storage?