To assess your Cloudbreak™ or Cloudbreak Freestyle™ sequencing run performance, start with reviewing the specifications for your AVITI™ or AVITI24™ instrument and compare them with the metrics reported for your run. You can review the following metrics easily from ElemBio™ Cloud:
Each of these metrics is described in more detail below, and you can also visit the ElemBio Cloud Sequencing Metrics page for more information.
Reads
The reads metric (shown in ElemBio Cloud in Figure 1) reports the number of polonies that have passed default filtering settings, unless filter mask settings are modified prior to the start of the run. This output from your sequencing run can be compared with the instrument specifications for your instrument/sequencing kit (linked above) to assess the performance of the run.
The amount of library loaded onto the sequencer has a major impact on the sequencing run output, so be sure to review library loading recommendations for your library size in the Cloudbreak Sequencing User Guide, and read the Knowledge Base article on Library Quantification and Loading for additional help. The pass-filter rate (PF) is also reported for your run, which shows the proportion of total sequencing reads that are passing default filtering settings.
The output that is predicted on the instrument screen during a sequencing run is an estimate of ~95% pass-filter reads and is updated after cycle 15 of Read 2 (for paired-end runs) as filtering settings are applied and the run continues. While these settings remove low quality reads based on performance in both read 1 and read 2 by default, they can be adjusted after the run completes with Bases2Fastq to shift filtering cycles, reduce the number of filtering cycles, or disable filtering completely. Changing or relaxing filtering settings may increase the number of pass-filter reads available for analysis at the expense of quality, so it is recommended that one applies custom filtering settings with careful consideration.
If filtering for your run is unusually high (resulting in a low PF), you may want to do some additional assessment to determine the cause. For example, you can recover reads that failed filtering settings by adjusting Bases2Fastq settings, allowing you to investigate the quality, base composition, and more of those reads. See the Knowledge Base articles on Pass-Filter Reads and Changing Settings with Bases2Fastq for more information on how to do this.
Thumbnail Image
The thumbnail image is a snapshot of the polonies on a tile from the first cycle of Read 1, and indicates qualitative sample density. Reviewing the thumbnail image along with the reads metric can help you identify under- or over-loaded sequencing runs. For example, an overloaded run will likely exhibit overly dense polonies in the thumbnail image, along with low pass-filter rates (PF), lower than expected read counts, and lower quality. In contrast, an underloaded run will exhibit less dense polonies, often with normal quality and pass-filter rates, but fewer reads than expected. See the example thumbnail images in Figure 2 for an underloaded, expected, and overloaded sequencing run.
Quality
You can view %Q30, %Q40, and average Q scores for your run in ElemBio Cloud under Primary Analysis. For UltraQ™ runs, you can also view %Q50. These metrics are expressed graphically by cycle for Read 1 and Read 2, and an average across the run is included for each read. In general, Cloudbreak and Cloudbreak Freestyle kit configurations on the AVITI and AVITI24 have specifications for quality that are greater than 85-90% Q30, depending on the kit type being used. UltraQ kit specifications include greater than 90% Q40 and greater than 70% Q50.
Quality often correlates with other metrics, so reviewing them together can provide additional insight into your sequencing run. While overall low quality for a sequencing run may correlate with overloading, low quality in specific regions or cycles can be associated with a low-diversity region of a library. You can confirm this by looking at the base composition metric and comparing this with your expected library structure. If you suspect that a low diversity region is in the beginning of your reads where filtering occurs by default, you can try changing filtering settings with Bases2Fastq using the –filter-mask flag and appropriate Base Masks to relax or prevent filtering in that region.
If you are sequencing a complex pool with multiple library types and sizes, you may want to investigate the quality of your reads across some of your samples. You can do this with the FASTQ files generated by Bases2Fastq and open-source tools like FastQC or MultiQC.
Base Composition
Base composition will vary with library structure, so this metric is useful to compare with your expected structure in mind. Generally, standard high-complexity libraries (like standard WGS and RNA-seq libraries) will have a roughly even base distribution across cycles through the library insert, with each base hovering around 20-30% with minor fluctuations. Libraries with regions of low nucleotide diversity (like amplicon libraries) will show a skewed base composition in areas where that low diversity occurs.
 |
Importantly for run performance, polony positions must be mapped in areas of high diversity for good general performance. Elevate libraries by default map polonies in a diverse region, but third party libraries require diversity in the first 5 cycles of read 1. Alternatively, you can use the Low-Diversity High Multiplexing setting if your library and indices meet the requirements, or spike in some amount of PhiX.
Additionally, you should look for evidence of “spikiness” in base composition after the indexing read. This may indicate the presence of adapter dimers and can also be associated with low quality. For more information on identifying adapter dimer in your library and how it can affect your sequencing run, see the Knowledge Base article on Library Quantification and Loading.
PhiX Error Rate
 |
For runs that include a PhiX control library spike-in, the PhiX Error Rate represents the percentage of bases per cycle that have errors relative to the PhiX control genome.
In general, PhiX in Cloudbreak sequencing runs should have low average error rates of around 0.2-0.5% per read. However, as PhiX tends to be a small proportion of the polonies on the flow cell, its error rates can be somewhat contingent on the overall flow cell density and performance of the other polonies. If PhiX error rates appear abnormal and the run doesn’t appear to be overloaded, please contact Element Biosciences Technical Support.
Index Assignment
 |
Index Assignment refers to the percentage of reads that are assigned to an index based on the sequences provided in the Run Manifest. Review index assignment to ensure all intended samples have been sequenced. If you see unexpected results or missing index sequences, be sure to check the Run Manifest first for any mistakes or missing samples. You can learn more about run manifest setup by reading the Run Manifest section in the Bases2Fastq documentation.