Stanford Diabetes Genomics and Analysis Core

Overview

The Diabetes Genomics Analysis Core (DGAC) at Stanford offers library preparation and sequencing services on a variety of platforms - Illumina HiSeq 4000, MiSeq, HiSeq 2500 and PacBio Sequel - as well as bioinformatics analysis. We can sequence a variety of commercial sample preparation kits as well as custom workflows. DGAC provides access to cost-effective high throughput sequencing and analysis to researchers at the Stanford Diabetes Research Center.

Services

Illumina Libraries

RNA-Seq Kits

• NGS Library Preparation
• mRNA
• Fluidigm C1 Single-Cell cDNA Libraries

10X Genomics

• Chromium Genome Solution
• Chromium Single Cell 3' Solution

Methyl-Seq

• Whole-Genome Bisulfite Sequencing

PacBio SMRTbell Libraries

SMRTbell library preparation does not utilize amplification steps; the completed library molecules are used as direct templates for the sequencing process. Due to this, the quality of the starting DNA has enormous influence on the success and quality of the sequencing results. High-quality, high-molecular-weight genomic DNA is extremely important for obtaining long reads and optimal performance.

Illumina Sequencing

HiSeq 4000

The Illumina HiSeq 4000 is over three times faster than an identical run length on the HiSeq 2000 while providing over 60% more reads per lane. However, the 4000 has more stringent library requirements and requires a larger sample size than the 2000.

A 2×101 run will take about 2-3 days to complete; a 2×151 will take 3-4. (Time variation can occur depending on the time of day clustering was performed: in the morning or afternoon).

The HiSeq 4000 generally gives 280-330 million reads per lane for a flowcell total of around 2.2 billion reads. While the patterned flow cell used on the 4000 is fixed at 480 million wells/reads per lane, Illumina recommends between 60-70% passed filter (PF) to maximize the number of unique reads in the lane. A higher PF will provide a higher amount of non-unique reads, which means more duplicate data. The HiSeq 4000 is biased towards smaller fragment sizes, so please try to keep your samples close to a consistent fragment length.

More information on the HiSeq 4000 specifications can be found on Illumina's website (PDF).

HiSeq 2500

The Illumina HiSeq 2500 allows more versatility in types of runs than the HiSeq 4000. As many clients are producing libraries that haven't been optimized or cleared for use on the 4000, the 2000/2500 are good substitutes. A 2x101 run will take about 11 days to complete; a 1x50 run will take 3 days (plus additional time for clustering). The HiSeq 2500 generally gives 190 million reads per lane for a flow cell total of around 1.5 billion reads.

More information on the HiSeq 2500 specifications can be found on Illumina's website (PDF).

Rapid Run Mode

Illumina indicates that the Rapid flowcells can get up to ~150 million paired reads per lane under optimal conditions. Rapid Run Mode is great for libraries that are not optimized for the 4000, require more data than a MiSeq provides, and need the results quicker than a 2500 high throughput mode will allow.

MiSeq

One of the great advantages to the MiSeq is the quick turnaround time for runs. The MiSeq is capable of long reads, making it great for de novo assembly of small genomes. The MiSeq is also great for QC tests on sequencing workflows before committing to larger batches on more expensive machines.

The MiSeq v2 kit provides 12-15 million single reads or 24-30 million paired-end reads, while the MiSeq v3 kit provides 22-25 million single reads or 44-50 million paired-end reads.

More information on the MiSeq specifications can be found on Illumina's website (PDF).

Pacific BioSciences Sequencing

Sequel

The new Pacific Biosciences Sequel system builds upon their Single-Molecule, Real-Time (SMRT) technology and delivers higher throughput, increased scalability and lower sequencing costs compared to the PacBio RS II.

SMRT Sequencing

The Sequel system utilizes the DNA sample as a direct template for the sequencing reaction in order to sequence individual DNA molecules in real time. Ligation of hairpin adapters converts the double-stranded DNA molecules into a circular template that enables continuous sequencing of both the forward and reverse strands without the need for amplification. The unique circular structure allows for continuous long reads of large-insert libraries or circular consensus sequencing of smaller insert sizes. Extra-long read lengths and high consensus accuracy make the Sequel a helpful tool for finding SNPs and sequencing through GC-rich areas of the genome.

Strengths and Applications

• Rapid, cost-effective generation of high-quality, de novo whole genome assembly
• No PCR bias, uniform coverage for targeted sequencing
• Direct base modification detection for epigenetic studies
• Long read lengths
• Single-molecule resolution of complex populations
• High consensus accuracy

Sequencing on the Sequel

The basic sequencing "unit" for the Sequel is the SMRTcell, analogous to a flowcell in Illumina sequencing. SMRTcells do not have lanes, unlike flowcells, so each sample is sequenced on its own SMRTcell. There are no "read lengths" like in Illumina sequencing; instead, you choose a length of time to continuously run the sequencing reaction, known as a "movie time". The maximum movie time for the Sequel is 10 hours.

There are two different methods of loading the SMRTcell: diffusion loading and Magbead loading. Fragment size largely determines which loading strategy you will use. For smaller fragment sizes (7.5kb), diffusion loading is generally the recommended method. In diffusion loading, SMRTbell library molecules with bound polymerase are immobilized and diffuse to the bottom of the ZMW wells after removal of excess polymerase and primer with cleanup beads. For fragments >7.5kb, Magbead loading is recommended, which removes adapter dimers and short insert sizes to maximize read lengths.

Bioinformatics Analysis

On-premise computational cluster: specially suited to NGS analysis

• 2800+ cores and 7+ Petabytes of high performance storage
• Architecture specifically suited to large scale genomics data analysis but also supports general scientific computing
• 500+ bioinformatics software packages installed and ready to use
• Specialized data analysis solution, Galaxy, available to users. Inquire for ongoing pilot programs.

"Bioinformatics-as-a-Service (BaaS)": consulting for hands-on help

• Payable on an hourly basis
• We not only support analysis for popular NGS data types such as RNA­Seq, ChIP­Seq, MethylSeq, Whole Genome/Whole Exome Seq, CancerSeq, and Microbiome, we also support new data types like Hi-C, and ATAC-Seq.
• Along with Secondary analysis, we provide consulting in quality control, downstream tertiary analysis, data interpretation and visualization. Custom/novel development is available.

Google Cloud Gateway: for scale out computing

• Cloud services backed by SU agreements
• Security compliance set up by DGAC system admins

System Security: supports biomedical research compliance needs

• Services support SU mandated security requirements e.g. two-factor authentication
• On-premise and Cloud gateway support. Learn more about dbGaP here.
• Peer reviewed publication on Cloud security
• Regular Stanford IT audits to keep up with security compliance requirements