HiSeq 3000

In July of 2015 the CGRB began offering high throughput sequencing on the Illumina HiSeq 3000 instrument. The HiSeq 3000 provides faster turnaround times, longer read lengths, and increased output compared to the HiSeq 2000 sequencing service previously offered. These improvements are due to the HiSeq 3000 having a faster camera and computing hardware, using patterned flow cells, exclusion amplification clustering, and faster sequencing by synthesis (SBS) reagents. Reads as long as 150bp paired end are available on the HiSeq 3000, which takes approximately 3.5 days to complete. Output is typically 300-400 million reads per lane or 90-120Gb per lane for a 150bp paired end run. Please note that a flow cell consists of 8 lanes, each of which must be filled before we can start a sequencing run. The more flexible you are with your run type, the quicker the turnaround will be. The CGRB offers library prep and QC services prior to sequencing. To help the Illumina HTS user community, we have established an HTS mailing list where users can share information and announcements are posted by the Core Lab. Please contact Mark Dasenko with any questions regarding library prep and sequencing submission. Questions regarding the sequencing data output can be addressed by Matthew Peterson.


Important information specific to the HiSeq 3000:  
1) It is not recommend to run libraries on the HiSeq 3000 with median fragment sizes >500bp, as this could lead to an increase in sequencing reads being filtered out. But, the CGRB has successfully run libraries with longer median fragment sizes with minimal impact on results.   

2) Adapter dimers can be more problematic on patterned flow cells (eg. 5% adapter dimer contamination can lead to 60% of the sequencing reads being adapter). It is highly recommended to check library sizing and adapter dimer content on the Agilent bioanalyzer or tapestation prior to sequencing.

3) Low diversity (at each cycle of sequencing, most bases are the same; base distribution not balanced) libraries can be problematic when run on any Illumina instrument, but particularly for the HiSeq 3000. Therefore, a control must be spiked into the sample by the Core Facilities prior to sequencing.  For low diversity libraries (amplicons, RAD-seq, GBS, etc.) it is recommended to spike in at least 10% PhiX. Failure to follow these guidelines can result in a failed sequencing run, in which the CGRB Core Facilities cannot be held responsible. It is most important to have high diversity during the first several cycles of sequencing, as this is when the critical calculations for the entire run are made (color registration, template generation, and normalization). 

4) Index hopping or switching is a known problem with Illumina platforms, but seems to occur more frequently with those that use exclusion amplification clustering, such as the HiSeq 3000.