F2 Support on ICA

Mandatory Upgrade for TSO on ICA

Amazon recently announced that their F1 instances, which are used by TSO 500 pipelines on ICA, will be replaced by a new generation of FPGA-powered cloud hardware, Amazon EC2 F2 instances. Furthermore, F1 instances will cease availability on December 20, 2025, their end-of-life date. To ensure uninterrupted service, Illumina is releasing new versions of current DRAGEN TSO 500 and DRAGEN TSO 500 ctDNA pipelines on ICA compatible with F2 instances.

F2-compatibility provides significant performance improvements: 40% lower turnaround time for TSO 500 analyses, on average. Furthermore, this suite of versions has been validated to produce bit-exact results.

To aide your transition to F2-compatible pipelines, this document contains:

Bit exact results comparisons

Bit-exact means that the new F2-compatible versions will have no analytical changes to the bioinformatics pipeline and results are identical to the current F1-compatible versions.

For each new pipeline version, the same data set was analyzed with both the F2-compatible and F1-compatible pipeline versions. For each output file in Logs_Intermediates/ and Results/ folders, F1 and F2 outputs were compared using an automated script. All differences were logged and manually reviewed.

Across all pipeline versions, the only changes observed were those that were expected: related to versioning and naming (pipeline name, analysis ID, time stamp, etc.) and the mapping rate, a metric found in mapping_metrics.csv, an output of DNA and RNA Map/ Align. Mapping Rate is a measure of speed, in millions of reads per second. It varies based on the hardware used, can be different between cloud, local, different fpga cards, etc. Changes to mapping rate do not impact map/ align outputs, and therefore there is no impact to results.

Turnaround time improvements

F2 instances introduce significant capacity and performance improvements compared to the currently used F1 instances, with approximately 40% lower turnaround time for the same TSO 500 analysis on F2 instances versus F1.

See a summary of turnaround time improvements in the table below. Run Time is the time analysis was ongoing, while Analysis Turnaround Time (TAT) is the end-to-end processing time: from submitting the analysis to results being available in ICA.

Pipeline Version
Run Time Improvement
Analysis TAT Improvement

DRAGEN TSO 500 v2.6.0.8

1.84x (46%)

1.69x (41%)

DRAGEN TSO 500 v2.5.2.6

2.46x (59%)

1.86x (46%)

DRAGEN TSO 500 ctDNA v2.6.1.8

1.55x (35%)

1.40x (29%)

DRAGEN TSO 500 ctDNA v2.6.0.25

1.90x (47%)

1.74x (42%)

TSO 500 Overall

1.80x (44% reduction)

1.65x (39% reduction)

Starting in May 2025, Illumina will be releasing F2-compatible versions (of some, but not all, current pipelines) in ICA.

For pipelines without bit-exact F2 versions, we recommend following Transition Path A and upgrading to the highest available F2-compatible pipeline (v2.6.0.8 tissue and v2.6.1.8 ctDNA).

For pipelines with bit-exact F2 versions, you may choose to transition to the bit-exact version (Transition Path B) or upgrade to the highest available version (Transition Path A).

For those following Transition Path A and upgrading their pipeline to a newer version, we provide additional information about changes between pipeline versions here.

Current Version
Recommended Version
Transition Path

v2.1.2

v2.6.0.8*

A: pipeline upgrade

v2.1.2 HRD

v2.6.0.8*

A: pipeline upgrade

v2.5.2

v2.5.2.6*

B: bit-exact

v2.5.2 HRD

v2.5.2.6*

B: bit-exact

v2.6.0

v2.6.0.8*

B: bit-exact

v2.6.0 HRD

v2.6.0.8*

B: bit-exact

*will include HRD

Current Version
Recommended Version
Transition Path

v2.1.1

v2.6.1.8

A: pipeline upgrade

v2.5.0

v2.6.1.8

A: pipeline upgrade

v2.6.0

v2.6.0.25

B: bit-exact

v2.6.1

v2.6.1.8

B: bit-exact

How to access new pipeline versions…

Access to new ICA pipelines depends on whether you launch analysis manually (via the ICA command line or user interface); or use auto-launch to automatically kick off analyses after sequencing.

…when manually launching in ICA

To kick off analysis with a new pipeline version, you need that pipeline to be accessible within your ICA project. In other words, the bundle containing that pipeline must be linked to your project.

If the new pipeline version is bit-exact to the current version, i.e. Transition Path B, there is nothing for you to do. The new version will be added to the same bundle as your current pipeline and automatically linked to your project. Simply select the new version from the list of pipelines available in your project when you want to use it for a new analysis.

If the new pipeline version is an upgrade from your current version, i.e. Transition Path A, you must link the bundle for that pipeline to your ICA project. The bundles are found in Entitled Bundles and named with the corresponding 3-digit pipeline version (e.g. DRAGEN TSO 500 v2.6.0). See the TSO 500 User Guide for instructions on getting started in ICA.

…when Auto-launching via BaseSpace Sequence Hub

To auto-launch a new pipeline version, you need to make sure your sample sheet has the correct URN. Updated URNs will be listed on this page when the F2-compatible pipelines are available. See the User Guide for instructions on Auto-launching TSO 500 pipelines.

Guidance for Upgrading TSO 500 Pipelines (Transition Path A)

The following sections contain information to guide users that are transitioning from one pipeline version to a higher one (e.g. 2.1 to 2.6). Analytical output changes are detailed along with differences in sample sheet and input requirements.

Bioinformatics changes

Refer to the tables below for a summary of analytical changes (algorithm updates, new variant types, accuracy improvements, etc.) between v2.1 and v2.6.

DRAGEN TSO 500 Analytical Changes

For a detailed understanding of analysis methodology and output files, please refer to the relevant section of the DRAGEN TSO 500 v2.6 user guide.

Analysis Step
v2.1.2 to v2.5.2
v2.5.2 to v2.6.0

Demultiplex

No change

No change

DNA Map/ Align

No change

No change

Small Variants

• Manifest expanded 8bp into introns

• MNVs and component SNVs/ indels all reported in VCF

• Accuracy improvements

No change

Tumor Mutational Burden

Updates from small variant caller

MNVs removed from TMB calculation

Microsatellite Instability

No change

No change

Copy Number Variants

• Report predicted sex

• Expand CNVs from 55 genes to 514

• Report deletions in addition to amplifications

No change

BRCA Large Rearrangements

• Breakpoint coordinates shifted (now 0-based instead of 1-based)

• Algorithm update for events with small fold change

Genomic Instability Score*

Accuracy improvements

No change

Loss of Heterozygosity*

New biomarker (beta feature)

No change

Absolute Copy Number*

New biomarker (beta feature)

No change

RNA Map/ Align

No change

Differences due to update in DRAGEN versions

RNA Fusions

No change

Accuracy improvements

Splice Variants

No change

Differences due to RNA map align updates

QC Metrics

Added: PER_GENE_MEDIAN_COVERAGE (RNA)

Added:

• gene- and exon-level coverage

• PCT_Q30

• PCT_SOFT_CLIPPED_BASES

Removed: PCT_PF_UQ_READS (%)

DRAGEN SW

v3.10.9 to v3.10.16

v3.10.16 to v3.10.17

*only for HRD samples

DRAGEN TSO 500 ctDNA Analytical Changes

For a detailed understanding of analysis methodology and output files, please refer to the relevant section of the DRAGEN TSO 500 ctDNA v2.6 user guide.

Analysis Step
v2.1.1 to v2.5.0
v2.5.0 to v2.6.0
v2.6.0 to v2.6.1

Demultiplex

No change

No change

No change

Map/ Align

No change

No change

No change

Small Variants

• MNVs and component SNVs/ indels all reported in VCF

• Accuracy improvements

No change

No change

Tumor Mutational Burden

No change

No change

No change

Microsatellite Instability

No change

No change

No change

Copy Number Variants

Report predicted sex

No change

No change

DNA Fusions

SV evidence BAM added to Logs_Intermediates/

No change

Improved SV detection in samples with high chimeric reads

QC Metrics

No change

• gene- and exon-level coverage

• PCT_Q30

• PCT_SOFT_CLIPPED_BASES

No change

DRAGEN SW

v3.10.9 to v3.10.15

v3.10.15 to v3.10.17

v3.10.17 to v3.10.18

Input changes

DRAGEN TSO 500 Input Changes

Refer to the table below for a summary of changes to input requirements between 2.1.2 and 2.6.0. To ensure your sample sheet will pass validation in the 2.6 pipeline, please refer to the sample sheet requirements detailed in the DRAGEN TSO 500 v2.6 user guide.

DRAGEN TSO 500 Inputs
v2.1.2
v2.5.2
v2.6.0

Sample sheet section: [Sequencing_Settings]

LibraryPrepKits field is now required.

Valid values are: TSO500, TSO500HT

Sample sheet section: [TSO500S_Data]

Index ID is now optional

FASTQ folder structure

Sub-folders per sample

Sub-folders per sample or one folder with all FASTQs

No change

DRAGEN TSO 500 ctDNA Input Changes

Refer to the table below for a summary of changes to input requirements between 2.1.1 and 2.6.0. To ensure your sample sheet will pass validation in the 2.6 pipeline, please refer to the sample sheet requirements detailed in the DRAGEN TSO 500 ctDNA v2.6 user guide.

Inputs
v2.1.1
v2.5.0
v2.6.0
v2.6.1

Sample sheet section: [TSO500L_Data]

Index ID is now optional

FASTQ folder structure

Sub-folders per sample

Sub-folders per sample or one folder with all FASTQs

No change

No change

References

For additional details on changes between versions, please refer to the customer release notes for not only the last version, but also each version in between.

Transition
Relevant Customer Release Notes

DRAGEN TSO 500 v2.1.2 to

DRAGEN TSO 500 v2.5.2

DRAGEN TSO 500 v2.5.2 to

DRAGEN TSO 500 v2.6.0

DRAGEN TSO 500 ctDNA v2.1.1 to

DRAGEN TSO 500 ctDNA v2.5.0

DRAGEN TSO 500 ctDNA v2.5.0 to

DRAGEN TSO 500 ctDNA v2.6.0

DRAGEN TSO 500 ctDNA v2.6.0 to

DRAGEN TSO 500 ctDNA v2.6.1

Last updated

Was this helpful?