The Illumina Protein Prep Automation System Output File (IPPAS output file) is a .csv file that's produced after automated library prep is completed. It contains the following fields:

• NGS readout of more than 9K unique protein targets for a single plasma or serum sample.

• From sample to processed results in under 2.5 days with just 4 hours of hands-on time.

• • Includes both local and cloud solutions for planning a run and for processing data with DRAGEN Protein Quantification.

E2E Overview

The Illumina E2E Protein Prep solution integrates automation steps to perform sample preparation, protein capture, sequencing, and bioinformatics analysis. Once sequencing is finished, the DRAGEN Protein Quantification application automatically initiates in BSSH or ICA. The diagram below illustrates the E2E workflow.

For documentation on the assay and automation components of Illumina Protein Prep, please refer to the Illumina Protein Prep Product Documentation (document # 200045446).

Normalization Steps

Normalization corrects for the sources of confounding variation, such as overall protein concentration differences, minor deviations in volume transfer during the assay, or efficiency of library preparation steps. It's performed sequentially and produces an individual ADAT file with counts for each normalization step. See the metrics appendix page for a summary of all metrics.

• Hybridization Normalization: This step corrects for biases that can occur during the hybridization and sequencing preparation stages of the assay. During hybridization, controls are spiked into each sample; during normalization, the counts for these controls are compared against an internal reference based on all the samples on the plate. A scale factor is calculated for each sample, and if the scale factor is outside of the specification range (0.4–2.5), the sample will receive a FLAG.

• Internal Reference Median Normalization: This step corrects for differences in the total protein abundance measurement of a sample. It is performed for each dilution group and runs separately for blank and calibrator control samples. A scale factor is calculated for each dilution group, by comparing the observed protein measurements to a reference of expected values for each protein.

  This reference is based on median protein counts across all samples of the same sample type on the same plate. If any of the scale factors for a sample are outside of the specification range (0.4–2.5), the sample will receive a FLAG.

• Plate Scaling: This step corrects for possible changes in measured total protein counts between plates, when calibrator samples are present. The median of each SOMAmer measurement across the five calibrators is compared to an external calibration reference to calculate a plate scale factor.

  The first plate scaling step compares the calibrator medians to a reference derived from the sequencing instrument (NovaSeq 6000 or NovaSeq X Series) and adjusts the entire plate accordingly.

  The second step compares the scaled calibrator medians to a reference derived from the NovaSeq X Series (10B flow cell). There is no metric for plate scaling.

  • The references used in this step can be found in the SOMAmer metadata, under Ref.Bridging.<CalibratorId>.<Instrument>.<Flowcell>.<MasterMixLot#>. The reference used by cross-instrument plate scaling is Ref.Bridging.<PlateBarcode>.NovaSeqX.10B.AA.

• Calibration: This step corrects for batch effects that impact individual SOMAmers.

  The first calibration step (Platform Specific Calibration) compares the median of each SOMAmer measurement across the five Calibrator sample replicates to an external Calibration reference. This reference is derived from runs using the same instrument type, flow cell type, calibrator lot, and sample input type used in the run being analyzed. It then calculates a SOMAmer-specific scale factor and a Calibrator metric (PlatformSpecificCalibrationTailPercent). PlatformSpecificCalibrationTailPercent corresponds to the percentage of SOMAmers with scale factors outside the specification range (0.6–1.4). If 15% of SOMAmers fall outside of this specification, the plate receives a WARNING for the PlatformSpecificTailPercent_PassFlag metric.

  The second calibration step (Cross Platform Calibration) compares the updated calibrator medians to a reference derived from the NovaSeq X Series (10B flow cell), using the same calibrator lot and sample input type as the run being analyzed. The scale factor used to align the median of the calibrators to the reference value is applied to all samples on the same plate.

  • The references used in this step can be found in the SOMAmer metadata, under Ref.Bridging.<CalibratorId>.<Instrument>.<Flowcell>.<MasterMixLot#>. The reference used by cross-instrument calibration is Ref.Bridging.<CalibratorId>.NovaSeqX.10B.AA.

• External Reference Median Normalization: This step corrects for differences in the total protein signal for samples on each dilution plate. A scale factor is calculated by comparing the observed protein measurements to a reference of expected values for each protein. It's performed on plasma/serum samples and QC samples.

  If any of the scale factors for a sample are outside of the specification range (0.4–2.5), the sample will receive a FLAG.

  • The reference used in this step can be found in the SOMAmer metadata, under Ref.MedNormExt.Plasma.QC or Ref.MedNormExt.Serum.QC (dependent on the input type).

Before setting up and running the Illumina Protein Prep End-to-End (E2E) solution, ensure that the necessary software, tools, and configurations are in place. These prerequisites can vary depending on the environment used to run the secondary analysis (e.g., via cloud or locally). Follow the steps below to configure instrument and software appropriately.

Instrument Software Prerequisites

Cloud Secondary Analysis Prerequisites

• BaseSpace Sequence Hub (BSSH) or Illumina Connected Analytics (ICA) account.

  • All BSSH and ICA subscriptions come with access to the DRAGEN Protein Quantification application.

  • [ICA subscribers] To view results in ICA, select ICA Run Storage in BSSH Workgroup Settings.

  • For more information about which subscription is best for your use case, contact your FAS or TAM.

Note: When performing analysis with more than one plate, each plate must have their own unique plate barcode value.

Local Secondary Analysis Prerequisites

• DRAGEN Phase 4 server with the following space recommendations:

  • NovaSeq6000, S4 flow cell: ≥ 10GB

  • NovaSeqX, 10B flow cell: ≥ 20 GB

  • NovaSeqX, 25B flow cell: ≥ 35 GB

• External storage drive mounted to a DRAGEN server. This drive must be mounted through a network share and support CIFS/SMB protocols. Read and write permissions are required to use this network share. For more information, please refer to the Illumina Protein Prep Product Documentation (document # 200045446).

Prerequisites

1. Local Proteomics Sample Sheet Generator Excel Workbook.

2. One or more IPPAS output files (one per plate).

3. Access to a DRAGEN Server configured to the Illumina Protein Quantification Local Secondary Analysis (see prerequisites page for more information).

Steps

1. Download and open the Local Proteomics Sample Sheet Generator excel file.

2. Navigate to the Start tab and follow the instructions described in the upper left portion of the sheet:

   1. Fill in the RunName and output_file_prefix.

   2. Change the InstrumentPlatform and PrepKitName as needed.

   3. Copy and paste values from IPPAS output file(s) not including the headers into the specified fields.

   4. Add the Flow Cell lanes for each sample as a comma separated list with no spaces as needed.

   5. Ensure there are no errors displayed in the Row Check or Overall Check (a green cell is expected)

   6. Navigate to the "SaveAsCSV" and save it as CSV (filename must not contain special characters).

   7. This is a sample sheet compatible to local DRAGEN Protein Quantification.

3. (Recommended) Load this sample sheet onto the sequencer prior to sequencing.

   1. If the sample sheet is not included prior to sequencing, the user must manually reference the sample sheet when running DRAGEN Protein Quantification later.

Proteomics Sample Sheet Generator

Lane Splitting

The purpose of lane splitting is to enable the reuse of sample indexes (barcodes) on the same flow cell. To accomplish this, the samples indexed with the same barcodes must be physically separated by placing them on different lanes of the flow cell.

Currently, lane splitting functionality in Illumina Protein Prep is only supported on the NovaSeqX platform. To use lane splitting, use the sample sheet to indicate which lane(s) each sample is found in. Recommendation: Edit the "Lanes" column of the Illumina Protein Prep Automation System output file(s):

1. Find the Lanes column in the Illumina Protein Prep Automation System output file.

2. Add lane numbers in which the sample is located as a comma-separated value. Example: '1,2,3,4' for a sample located in lanes 1, 2, 3, and 4.

Multi-Analysis by Project

The goal of multi-analysis by project is to improve flexibility by enabling multiple analysis outputs from a single sequencing run without needing to requeue the sample sheet. A single project must include at least one plate (with its controls) and can include multiple plates used in a sequencing run. Each project produces one set of output files (including normalized ADATs and a DRAGEN Report).

To use multi-analysis by project, include the project name in the Project column of the Illumina Protein Prep Automation System output files. If no project information is provided, DRAGEN Protein Quantification will assume all plates belong to the same project and will assign the project name based on the user provided 'run name'. The multi-analysis by project feature is available for both the NovaSeq 6000 and NovaSeq X platforms.

1. Find the Project column in the Illumina Protein Prep Automation System output file.

2. Edit Project column in an Illumina Protein Prep Automation System output file by assigning a project name to each sample. Project name string should contain only alphanumeric characters and underscores.

3. Upload modified Illumina Protein Prep Automation System output file to the Run Planner.

Note: If you want all samples from a sequencing run to be analyzed together, DO NOT include any values in the "Project" column of the sample sheet.

Note: If project splitting is enabled, all samples on one plate must be included in the same project. Samples from one plate CANNOT be split across projects.

Lane Splitting and Multi-Analysis Examples

Here is an example of what the first lines of two individual IPPAS output files may look like after project and lanes values were added. The plate with Barcode "A" adds lanes "1,2,3,4" and project value "ProjectA". The plate with barcode "B" adds lanes "5,6,7,8" and project value "ProjectB". Note that, this configuration indicates that pro

Sample Quality

The purpose of a flag is to highlight that a sample required a high degree of correction during the normalization process. This means a sample had sufficiently high or low signal, causing the normalization scale factors to be out of specification. The general recommendation is to exercise caution when using that sample in downstream analysis; normalization may not have been able to properly correct for the large changes in signal.

Flag at SOMAmer Reads

A flag in a non-blank sample indicates low SOMAmer read depth. DRAGEN Protein Quantification has a minimum read depth to ensure measurement precision is achieved for each sample. Flagged samples may have a decrease in measurement precision.

Flag at SOMAmer Normalized Reads

A flag in a blank sample may indicate increased background or contamination.

Flag at Reference Correlation

A flag in a blank sample may indicate plasma or serum contamination. In internal studies, uncontaminated blanks generally have low RefCorr values (~0.4) and blanks with greater than 2% plasma or serum contamination were observed to have a RefCorr values of at least 0.6.

Flag at Hybridization Normalization

A flag in a sample indicates that the hybridization controls in that sample had elevated or decreased signal compared to the plate reference. Potential reasons for a high or low hybridization control signal include:

• Issues with the sample itself. These include poor sample quality, protein loss during hybridization (e.g., due to elevated temperature) or other issues that arise during library prep. In this case, other failure modes will likely also be flagged, such as low Raw Counts, or high / low MedNormExt scale factors.

• High plate-wide plasma / serum variability. Because Hyb Norm is performed using an internal reference, the composition of the plate can impact the scale factors of other samples. Therefore, a flag in one sample may reflect the quality of the other samples on the plate rather than issues with the sample itself (for example, many hemolyzed samples or blanks). In this case, other error modes for that sample are not expected to be flagged.

Flag at Internal Reference Median Normalization

A flag in a calibrator indicates that the flagged calibrator replicate had a large discrepancy in signal when compared to the median of all calibrators within that plate. A single flagged calibrator will have limited impact on the plate due to use of median calibrator values in analysis. Multiple flagged calibrators could impact plate performance; reach out to Illumina Support for additional information.

Flag at External Reference Median Normalization

A flag in a sample indicates that it had a large discrepancy in signal when compared to the external plasma or serum reference. Low signal may be caused by sample degradation or sample dilution; high signal may be caused by hemolysis. A difference in signal could also be caused by the diseased-state of the sample.

Plate Quality

QC Percent in Tails and Calibration Percent in Tails are used to determine plate quality by examining individual SOMAmers in each well. A QC Percent in Tails failure will likely require a repeat library preparation or re-sequencing for this plate; contact Illumina Support for additional information.

Review the following table for more details.

Plate Quality Matrix

Run Quality

There are two additional metrics on run quality that evaluates blank samples in a plate (see below). The general recommendation is that, if warnings are observed in one of these metrics, the plate should be used with caution in downstream analysis since there may have been sample contamination or elevated background. If issues with run quality are consistently seen, there may be issues with the assay set up. Contact Illumina Support for additional information or support.

Warning for SOMAmer Normalized Reads

A warning on a plate may indicate plate-wide increased background or contamination.

Warning at Reference Correlation

A warning on a plate may indicate plate-wide plasma or serum contamination.

QC Checks

There are a number of quality control checks that are applied on a plate and sample level. See the metrics appendix page for a summary of all metrics.

• Minimum SOMAmer Read Counts: Non-blank samples with less than 10 million reads will receive a FLAG for SOMAmerReads_PassFlag in the ADAT. These reads are counted in the raw counts step. Only human protein SOMAmers are part of this count, not controls. There is no specification for blank samples.

• Maximum SOMAmer Read Counts: Blank samples with more than 40 million normalized reads will receive a FLAG for SOMAmerNormRead_PassFlag in the ADAT. These reads are counted in the plate scale normalization step. Only human protein SOMAmers are part of this count, not controls. There is no specification for non-blank samples. A plate where 70% of the blanks have a FLAG for this step will receive a WARNING.

• Reference Correlation: This step produces a Spearman correlation coefficient describing how similar a sample is to a an external Plasma or Serum reference (see below). Blank samples with a correlation to the reference greater than 0.6 will receive a FLAG for RefCorr_PassFlag in the ADAT. A plate where 70% of the blanks have a FLAG for this step will receive a WARNING.

  • The reference used in this step can be found in the SOMAmer metadata, under Ref.MedNormExt.Plasma.QC or Ref.MedNormExt.Serum.QC (dependent on the input type).

• QC Check: This step compares the median of each SOMAmer measurement, across the three QC sample replicates, to an external QC reference. It then calculates a SOMAmer-specific scale factor and a QC metric (QCCheckTailPercent). QCCheckTailPercent corresponds to the percentage of SOMAmers with scale factors outside the specification range (0.8–1.2). If more than 15% of the scale factors are outside of the specification range, the plate receives a FAIL.

  • The references used in this step can be found in the SOMAmer metadata, under Ref.QCCheck.Plasma or Ref.QCCheck.Serum (dependent on the input type).

DRAGEN Protein Quantification Application

The DRAGEN Protein Quantification application is designed to perform counting and normalization for proteomics data from the Illumina Protein Prep pipeline. It converts data from the binary base call (BCL) files, generated by Illumina NovaSeq 6000 or NovaSeq X Series systems, into the normalized proteomic counts. Upon completion of sequencing, the application is automatically initiated for analysis on BaseSpace Sequence Hub (BSSH) or Illumina Connected Analytics (ICA).

The sections below exemplify how to configure the instrument(s) for autolaunching the secondary analysis, manually initiate an analysis, and requeue an analysis.

Setting Up Autolaunch on Sequencing Instrument

1. On your instrument, log into your workgroup and select the following run setup options:

   1. Workflow: NovaSeq Standard

   2. Read length: 15, 10, 10, 0

2. [NovaSeq 6000] Upload the v2 sample sheet generated by the BSSH Run Planner tool to the sequencing instrument.

3. [NovaSeq X Series] No action is required. The v2 sample sheet automatically displays on the instrument as a planned run.

4. Select the appropriate Illumina Protein Prep custom recipe for your sequencing instrument and flowcell. During the run, data is uploaded automatically to BSSH. Primary analysis and secondary analysis are completed automatically through autolaunch.

5. [Optional] Use the BaseSpace Sequence Hub to monitor the run from start to completion.

Manually Kicking Off Secondary Analysis

If manual mode sequencing is performed, or autolaunch is unsuccessful, there is an option to manually upload a completed run folder to BSSH and kick off the autolaunch analysis. To use this method, the samplesheet must be in the uploaded run folder, and be named "SampleSheet.csv".

Requeue

Sample Sheet Fields

This is a non comprehensive list of fields.

Samplesheet Examples

Examples of local and cloud sample sheets for NovaSeq 6000 and NovaSeq X are attached to this page.

Local

Cloud

DRAGEN Protein Quantification can be run locally after the installation of the local solution on a DRAGEN phase 4 server by an FAS.

To initiate a run:

The --analysisFolder parameter is optional. If no path is provided, output files will be put in a subdirectory of the folder where the command is being run.

The -s parameter is optional if the sample sheet file is included in the run folder.

For details on the parameters used with the script, execute the following command.

Primary Metrics

To view primary metrics:

1. Go to the relevant BaseSpace Sequence Hub (BSSH) workgroup.

2. Navigate to the "Runs" tab and select the run.

3. The "Summary" tab gives an overview of the sequencing run quality (e.g., average %Q30, %PF Yield).

4. Navigate to "Metrics" for detailed per lane information on all sequencing metrics.

Secondary Results

To view the analysis associated with a specific sequencing run:

1. Navigate to the "Summary" tab of the relevant run.

2. Click on the link below "Latest Analysis" (displaying the results from the latest analysis processed to the run data). For re-queued/re-analyzed runs, the previously completed analyses can be found under the "Prior Analysis" section

3. Click on "Reports" and find the quality metrics associated with the secondary analysis on the run data.

Note: Those who have access to an ICA account and want to view results on ICA may either click on "View Files in ICA" in the top right corner of your BSSH Analysis page or directly access the analysis in ICA. The secondary analysis results in ICA will be in a BSSH-managed project with the same name as the BSSH workgroup where the analysis was performed.

DRAGEN Reports is an HTML report that provides a quick overview of the quality of an E2E analysis. The report consists of three sections, which are displayed as tabs in the report. The following sections describe the tabs.

Plate QC

This section is subdivided into the following subsections:

• Reagent Lot Summary: This table describes the reagents used per plate in the analysis.

• Plate QC Summary: This table provides the plate level metrics, including Calibration % in Tails, QC % in Tails, Reference Correlation, and Blank Background metrics.

• Calibration Scale Factors: This histogram illustrates the distribution of calibration scale factors for a given plate.

• QC Scale Factors: This histogram illustrates the distribution of QC scale factors for a given plate.

Sample QC

This section of the report contains information on which samples passed or flagged specifications as well as SOMAmer count yield per sample. It is subdivided into the following subsections:

• Sample QC Summary: The table describes the percentage of samples (organized by sample type) that passed Quality Control.

• QC Summary: The heatmap shows the QC status of samples based on their position in the plate wells.

• Flagged Samples: The table identifies samples that failed one or more sequencing or normalization specifications.

• SOMAmer Read Counts: This graph represents the SOMAmer count for each sample included in the analysis.

Specification

This report section details the QC metrics and normalization steps.

DRAGEN Protein Quantification produces the following key output files in BaseSpace Sequence Hub:

• DRAGEN_Protein_Quantification_<SW Version>

  • <project> (if no projects are provided, there shall be one folder titled with the RunName)

    • adat

      • <output file prefix>_FinalNormStep_MedNormExt.adat: This file contains the final normalized counts for the samples in this project.

      • other_normalization_steps: This folder contains all intermediate normalized ADATs with their respective counts.

    • quality metrics

      • <output file prefix>_run_qc_stats.csv: This file provides a summary of per sample run qc statistics.

    • DRAGEN Report

• DRAGEN Report: Folder containing all projects DRAGEN report file/s

• Extra Files

• ICA Logs: Folder containing information on ICA logs

Example of Output Structure in BSSH

Sample Level Metrics

Sample level metrics are found in the sample/row-level metadata in the ADAT.

SOMAmer Level Metrics

SOMAmer level metrics are found in the SOMAmer/column-level metadata in the ADAT.

These PASS/FAIL designations are only used for plate level quality control. They should not be used to assess SOMAmer quality.

Plate Level Metrics

Plate level metrics are found in the header metadata in the ADAT.

ADATs can be analyzed in R or Python using parsers created by Somalogic.

DRAGEN Protein Quantification v2.0.0 is compatible with SomaData v1.0.0 (Python - formerly called Canopy) and SomaDataIO v6.1.0 (R).

Somadata creates an ADAT object, which is an extension of a Pandas DataFrame.

• rows correspond to samples

• columns correspond to SOMAmers

• values are normalized counts

Below are examples on parsing an ADAT in Python and R.

Parsing ADAT in Python

import somadata

# read the adat
my_adat = somadata.read_adat('/path/to/my/file1.adat')

# retrieve sample metadata
sample_meta = my_adat.index.to_frame(index=False)

# retrieve SOMAmer metadata
soma_meta = my_adat.columns.to_frame(index=False)

# retrieve the scale factors for all plates
plate_scale_factors_dict = my_adat.header_metadata['PlatformSpecificPlateScale_ScaleFactor']

### additional optional manipulation 

# concatenate multiple adats into a single adat
my_adat2 = somadata.read_adat('/path/to/my/file2.adat')
my_merged_adat = soma.smart_adat_concatenation([my_adat, my_adat2])

# write it to a file 
my_merged_adat = my_merged_adat.to_adat('/path/to/merged/adat')

Parsing ADAT in R

library(SomaDataIO)

# check all package functions
ls("package:SomaDataIO")

#read the adat
my_adat <- read_adat('/path/to/my/file1.adat')

# retrieve sample metadata
sample_meta <- my_adat[getMeta(my_adat)]

# retrieve SOMAmer metadata
soma_meta <- my_adat[getAnalytes(my_adat)]

# create a function that parses the header
parse_adat_header <- function(adat_file, max_lines = 500) {
  header_lines <- readLines(adat_file, n = max_lines)
  table_row <- grep("TABLE_BEGIN", header_lines)
  header_lines[1:(table_row-1)] %>% 
    strsplit("\t") %>% 
    {
      values <- map(., 2) # Get values from key value pairs
      keys <- map_chr(., 1) # Get keys...
      setNames(values, keys)
    }
}

my_adat_header <- parse_adat_header('/path/to/my/file1.adat')

# retrieve the scale factors for all plates
plate_scale_factors_dict = my_adat_header$PlatformSpecificPlateScale_ScaleFactor

### additional optional manipulation 
my_adat2 = read_adat('/path/to/my/file2.adat')
my_merged_adat = rbind(my_adat, my_adat2)

# write it to a file 
my_merged_adat = write_adat(my_merged_adat, '/path/to/merged/adat')

Illumina Connected Multiomics provides interactive visualizations and powerful statistics.

• For product information, please reference the ICM product documentation

This is a walkthrough of an analysis that could be done on ICM with an example proteomic data set, produced by DRAGEN Protein Quantification. It covers the following features:

• Adding samples to a study

• Filtering samples

• Data Transformation

• t-SNE

• PCA

• Hierarchical clustering and creating heatmaps

• Differential expression

• Gene set enrichment analysis

• Create two sample groups - one for female samples, and one for male samples. Filter by 'Sex', and select 'F' only. At the top of the page, click on ‘+ Create sample group’ to create a sample group.

• Type in a group name and click on the ‘Create’ button. Then, change the filter from 'F' to 'M', and create another sample group.

  • Navigate to the ‘Sample Group’ tab and confirm the sample groups.

Create an Advanced Analysis

• Click on ‘+ New Analysis’ to create a new analysis.

• In the pop-up window, provide a name for the analysis, select ‘Advanced Analysis’ as the Analysis Type, choose the sample group to be included in the analysis ('All ADAT Samples' will be selected by default), and click on the ‘Run Analysis’ button.

• Note: make sure there are no duplicated Sample IDs in the analysis groups.

• A pop-up message will show up if the analysis creation is successful.

• Refresh the page to get the latest status of the analysis.

• When the Status is ‘Complete’, click on the analysis tile to enter the analysis module.

• Click on the ‘Sex in F, M’ node generated from the previous step; select ‘Normalization and scaling’ > ‘Normalization’ in the right-hand-side Quantification toolbar.

• In the normalization view, drag and drop ‘Add’ and ‘Log’ from the ‘Available methods’ column to the 'Selected methods' column. Select 10.0 in the dropdown menu for ‘Log’, and click on the ‘Finish’ button.

PCA

• Click on the ‘Normalized counts’ node generated from the data transformation step, and select ‘Exploratory analysis’ > ‘PCA’ in the right-hand-side toolbar.

• In the PCA view, adjust the corresponding PCA parameters and click on the ‘Finish’ button.
• Compute biomarkers contributing to sex variable:

  • Click on the ‘PCA’ node generated from the previous step, then select ‘Statistics’ > ‘Compute biomarkers’.

  • In the compute biomarkers view, select corresponding parameters and click on the ‘Finish’ button. Double click on 'Biomarkers' to explore further.

t-SNE

• Click on the ‘PCA’ node generated from the previous step, and select ‘Exploratory analysis’ > ‘t-SNE’ in the right-hand-side toolbar.

• In the t-SNE view, select corresponding parameters and click on the ‘Finish’ button. Select the t-SNE node for additional information.

Differential Expression

• Click on the ‘Normalized counts’ node generated from the data transformation step. In the right-hand-side toolbar, select ‘Statistics’ > ‘Differential analysis’.

• In the differential analysis view, select a method (using Limma-trend in this example) and click on the ‘Next’ button.

• Select the factors to be included in the model and click on the ‘Add factors’ button. Confirm the selected factors show up in the ‘Selected factor(s)’ section, and proceed to the ‘Next’ button.

• Note: For paired or longitudinal analysis, check ‘Random’ column for certain variables to make them a random effect in the mixed effect model.

• If a categorical factor is selected (for example, 'Sex'), the next step will be to 'Define comparisons'. Add ‘F’ to the Numerator box and add ‘M’ to the Denominator box. Click ‘Add comparison’ button to create this comparison.

• The specified comparisons will appear in the ‘Comparisons’ section. Click on the ‘Finish’ button.

Filtering significantly differential analysis results

• Filtering significant differential analysis results

  • Click on the ‘F vs M’ node generated from the differential analysis step, and in the right-hand-side of toolbar, select ‘Filtering’ > ‘Differential analysis filter’.

  • Select ‘Metadata’ as the filter type, and specify the filter criteria. In this example, include features with false discovery rate (FDR) less than 0.05 for the Female vs Male comparison. Double click the filtered feature list node to view the Limma-trend report.

Hierarchical clustering / heatmap

• Click on the ‘Filtered feature list’ node generated from the filtering differential analysis results step, and select ‘Exploratory analysis’ > ‘Hierarchical clustering / heatmap’ from the right-hand-side toolbar.

• In the Hierarchical clustering / heatmap view, select ‘Assign order’ by Sex in the Sample order section, and click on the ‘Finish’ button.

• Double click on the ‘Hierarchical clustering / heatmap’ node generated from the previous step to view the heatmap with hierarchical clustering.

• In the left-hand-side toolbar, select ‘Annotations’ and add 'Sex' as a row annotation.

• View the heatmap - two major clusters can be observed: a cluster of protein signals with higher abundance in females compared to males, and a cluster with higher abundance in males compared to females.

• Select ‘Save as’ in the left-hand-side toolbar to save the heatmap to the data viewer.

Gene set enrichment analysis and pathway analysis

• Click on the ‘Filtered feature list’ node generated from the filtering differential analysis results step, and select ‘Biological interpretation’ > ‘Gene set enrichment’ in the right-hand-side toolbar.

• Select ‘KEGG data base’ in the Database section and select a corresponding library version. If none are available, click "New library" and the respective species. Then, click on the ‘Finish’ button.

Process FAQs

• How can I share data with a collaborator in BSSH?
• I sequenced in manual mode/need to kick of autolaunch after sequencing is complete. How can I do this?

  • Use the BSSH CLI to upload the run folder to BSSH. Make sure the samplesheet is named "SampleSheet.csv" and you are using an up-to-date version of the tool - at least v1.6.1.

  • For additional information, see the following documentation:
• How do I download/export the samplesheet?

  • Go to the BSSH Run Planner:

    • Go to BSSH, select the intended workgroup on the top right, and then select the Runs tab

  • For a new run:

    • Select the "New Run" button on the right and plan a new run with the desired pipeline.

    • After ingesting the IPPAS output files in the BSSH Run Planner, the user will reach the following Run Review page:

    • Select the "Export" option on the bottom right to download the samplesheet created

    • Select the "Save as Draft" or "Save as Planned" options to save this run as a draft or planned run respectively\

  • For an "Active" or "Planned" run:

    • Select the intended tab mentioned under the Runs header.

    • Select the required run from the list of runs:

    • Next, click File > Download > Sample Sheet:
• How can I requeue a run on BSSH with a new samplesheet/new version of the pipeline?

  • Go to the BSSH Run Planner tool and plan a new run with the desired pipeline.

  • Save the exported samplesheet (see the above step for additional information).

  • Go to the original run, and click Status > Requeue > Planned Run.

  • Select "Use a new Sample Sheet".

  • Upload the new sample sheet.

  • On the next page, click "requeue".

• Can multi-analysis by project split samples on the same plate?

  • No. All samples on the same plate must have the same project.

Metrics and Results FAQs

• What is DRC_Level and why is this used?

  • DRC stands for Dynamic Range Compression. It is used to even out the SOMAmer concentrations from a 5-log dynamic range to 2-log. Without DRC, the most abundant SOMAmers would occupy the majority of sequencing readout capacity. Each sample would require extremely high sequencing depth to cover the unabundant SOMAmers and accurately quantify them. In using DRC, probes are grouped based on their observed abundances under no DRC condition, and each group is compressed by a different set ratio. The compression level of each SOMAmer is included in the DRC_Level row of SOMAmer metadata.

• What is the LoD, and what does it indicate?

  • LoD stands for Limit of Detection. Each SOMAmer has an LoD for both Plasma and Serum. The LoD is a noise floor for that SOMAmer's fully normalized (MedNormExt) counts. In other words, normalized counts above the SOMAmer's LoD reflect the relative abundance of that SOMAmer's target protein, whereas normalized counts below its LoD may reflect assay uncertainty.

• How should LoD be used to discern signal?

  • The normalized count of a given protein is considered signal if it is above the LoD for that protein for it's given sample type.

• How do Somalogic and Illumina metrics compare?

  • Metrics and Normalization steps with different names between Somalogic's tool and Illumina's DRAGEN Protein Quantification. See the table below for the approximate mapping.

• Where are the FASTQs?

  • DRAGEN Protein Quantification does not produce FASTQs. By utilizing DRAGEN Counting, the software uses BCL Convert to demultiplex and count proteins per sample at the same time. This reduces analysis time but also removes FASTQ as a pipeline output.

• What if the PF or Q30 values are lower than expected?

  • First, check secondary metrics. If secondary metrics are passing, it's acceptable to proceed with further analysis. If secondary metrics have warnings or failures, consider re-pooling or repeating the dilute and denature and sequencing of the saved pool. If the data still looks poor, reach out to tech support for further guidance.

• What organisms does the SOMAmer metadata cover in the 9.5k assay?

The primary output file of DRAGEN Protein Quantification is the ADAT. One ADAT is produced for each normalization step performed. The final ADAT produced (<output_file_prefix>_Step7_FinalNormStep_MedNormExt.adat) has the normalized counts at the end of the full normalization process.

There are four key components to the ADAT: the header, SOMAmer metadata, sample metadata, and counts.

ADAT Header

Sample Metadata

SOMAmer Metadata