SOMAscan Assay FAQs

Frequently asked questions about the SOMAscan assay

Please choose a category below:

SOMAscan Assay-Specific Questions

What analytes are measured in the SOMAscan 1.3k Assay, and are they specific to a certain disease area?

You can download the list of over 1.3k analytes here or by emailing a request to information@somalogic.com. The SOMAmer reagents in the assay cover a broad array of proteins associated with cellular processes and disease patho-physiology as shown in the bar chart below. However, please note that this information relies on published scientific literature; for example, because of the emphasis on cancer research, more proteins have been measured and reported in that area than any other. If a particular protein has not been measured in a specific area or it has not been reported in the literature it will be “silent” rather than truly absent.

What controls are used in the SOMAscan assay?

We use both internal calibrator samples that are pooled from multiple individuals and quality control samples that are from a defined population. Calibrator samples are used as a reference standard. Internal quality control samples are also included to ensure that previously measured differences between calibrator and quality control samples are maintained. Spiked analytes are currently not used as a reference in the SOMAscan Assay, although we do this with our smaller panel assays. If the calibrators and QC samples fail, then we assess the assay as a failure.

Are samples assayed "neat"?

Plasma and serum samples are divided into three different dilutions and assigned a set of SOMAmer reagents to each dilution based on empirically detected levels in healthy individuals for that biological matrix. The most abundant proteins are targeted in the 0.005% sample dilution with the 0.005% SOMAmer reagent mix, and the lowest abundance proteins are targeted in the 40% sample dilution with the 40% SOMAmer reagent mix.
Non-plasma samples may be divided into different dilutions specific for that matrix. For example, the detection range for a single analyte in human plasma may be at a dilution of 40%, while the same analyte might have an optimized detection range in human CSF at a dilution of 1%.

Why are so many intracellular proteins detected?

In our internal studies, we have found proteins thought to be “intracellular” are present in plasma and serum. It is not uncommon to observe this event. There is a growing body of evidence, and with it heightened interest in, what researchers are calling the “exoproteome.” The exoproteome is defined as a group of proteins that appear in the fluid proximal to a given cellular system and arise from secretion, export, lysis and/or death. Below is a list of several recent publications that outline current thinking and examples of exoproteome research. It is worth noting that this subset of the proteome has only become of real interest as analytical methodologies, such as the SOMAscan Assay, have become more sensitive and reliable in detecting proteins at low levels.
References:

  • Armengaud, J, Christie-Oleza JA, Clair G, et al. Exoproteomics: Exploring the World Around Biological Systems. Expert Rev. Proteomics. 2012; 9(5):561-575.
  • Farrah T, Deutsch EW, Omenn GS, et al. A High-Confidence Human Plasma Proteome Reference Set with Estimated Concentrations in PeptideAtlas. Mol. Cell. Proteomics.  2011; 10(9): M110.006353. DOIL 10.1074/mcp.M110.006353.
  • Hortin G, Sviridov D. The Dynamic Range Problem in the Analysis of the Plasma Proteome. Journal of Proteomics. 2010; 73:629-636.
  • Karagiannis GS, Pavlou MP, Saraon P, et al. In-depth Proteomic Delineation of the Colorectal Cancer Exoproteome: Mechanistic Insight and Identification of Potential Biomarkers. Journal of Proteomics. 2014; 103:121-136.
  • Perkins, G, et al. Multi-Purpose Utility of Circulating Plasma DNA Testing in Patients with Advanced Cancers. PLoS One. 2012; 7(11): e47020.
  • Schwarzenbach H, Hoon DS, Pantel K. Cell-free nucleic acids as biomarkers in cancer patients. Nat Rev Cancer. 2011; 11: 426–437. doi: 10.1038/nrc3066.
  • Thierry AR, Mouliere F, Gongora C, Ollier J, Robert B, et al. Origin and quantification of circulating DNA in mice with human colorectal cancer xenografts. Nucleic Acids Res.  2010; 38: 6159–6175. doi: 10.1093/nar/gkq421.
  • Van Summaren A, Renes J, van Delft JHM, et al. Proteomics in the search for mechanisms and biomarkers of drug-induced hepatotoxicity. Toxicology in Vitro. 2012; 26:373-385.
What is the standard turnaround time for a study?

The standard turn-around time (TAT) for a study is 5 weeks from contract execution and sample receipt to data delivery for most studies (very large studies could take longer). Historically, our average TAT is 34 days. In certain situations, SomaLogic may be able to expedite studies if there are pressing time constraints; please contact your Sales Representative.

How is non-specific binding minimized in the SOMAscan assay?

SOMAmer reagents were designed to have long half-lives (i.e., slow off-rates). Specific interactions between SOMAmer reagents and proteins remain intact while non-specific binding events – which have quick off-rates – are disrupted.
The first interaction is when the sample competes for reagent binding during equilibration and the second is when the protein-SOMAmer reagent complex is released back into solution by photocleavage to remove the complex from the streptavidin bead. Kinetic challenge occurs during this photo-cleavage step. The SOMAscan assay takes advantage of the slow off-rates of the SOMAmer reagents by including excess competitor in this photo-cleavage solution; quickly dissociating complexes are prevented from rebinding and specific complexes are selectively enriched.

SOMAscan Assay Matrix-Specific Questions

Can samples from non-human species be assayed with the SOMAscan assay (e.g., mouse or rat)?

The SOMAscan assay can reproducibly and reliably measure hundreds of proteins in small sample volumes from mice, rats, cats and dogs better than any available alternative; non-human primates appear to cross-react even more. Using the SOMAscan assay to measure the proteome of non-human species has already led to novel discoveries indicating the SOMAscan assay can offer powerful, unbiased proteomic discoveries leading to deeper biological understanding of these species. Utilizing the same reagents in multiple species and in in vitro studies, creates an opportunity for a single translational platform, minimizing the technical risk of moving between different platforms.
Each matrix/species is optimized for overall performance of the assay; however at this time, we do not have species-specific panels. The volume of sample required depends upon the number of dilutions (1-3 dilutions per matrix) and the highest concentration for each dilution: the higher the concentration, the more volume is needed.

 

How does the SOMAscan Assay correlate with immunoassays (ELISAs)?

Based on a small number of comparisons of the SOMAscan Assay to ELISAs, we believe the SOMAscan Assay correlates with ELISAs comparably to the degree that ELISAs correlate with other ELISAs.    Please see our Technical Note Correlation of SOMAmer® Reagents in the SOMAscan® Assay and Commercially Available Immunoassays for additional details.

 

Can the SOMAscan assay be run with my matrix?

We request that you provide the details about your matrix, including the sample collection and processing protocols, for our technical specialists to review. Sample preparation must preserve the native, folded structure of the protein and the sample should not contain DNA-binding substances. During this review our team will look for substances that may interfere with the assay. After the feasibility review, it may be possible that we can assay your matrix according to our already defined protocols or we may need to discuss a new matrix feasibility assessment.

 

SOMAmer Reagent Questions (specific to SOMAscan assay)

How did SomaLogic select the SOMAmer reagents for the SOMAscan assay?

We selected the SOMAmer reagents based upon the following criteria:
1) Proteins that were thought to be present in human plasma or serum, or expressed in a wide variety of conditions.
2) Commercially available human proteins, whether or not they were known to be expressed in plasma or serum.
3) A SOMAmer reagent identified that bound with Kd < 10 nM and passed subsequent QC criteria. Larger Kd values may be permitted for highly abundant analytes.

 

Can the SOMAmer reagents selectively bind proteins that have post-translational modifications (PTM)?

SOMAmer reagents bind proteins in their folded forms. If a post-translational modification alters the structure of a protein or occludes the binding of a SOMAmer reagent, we may be able to generate SOMAmer reagents that differentiate between the different protein forms. However, in the absence of a structural change, a SOMAmer reagent will likely bind both forms equivalently.

 

Can SOMAmer reagents bind small proteins/polypeptides?

SOMAmer reagents bind natively folded proteins. It is challenging to generate SOMAmer binding reagents to the shorter polypeptides because they are typically more flexible and may not hold a stable structure. The more flexible the protein the more challenging it is to generate SOMAmer reagents because our SOMAmer reagents recognize stable tertiary structures. Within the SOMAscan assay menu there are SOMAmer reagents that bind to small polypeptides, in the <5 kDa range, such as insulin, b-endorphin, BNP-32, glucagon, etc.

 

How are the SOMAmer reagents characterized before inclusion into the SOMAscan assay menu?

Reagents are qualified for use in the SOMAscan assay if the apparent Kd for the selected protein is 1×10-9 M or better, the RFU range is > 1.5 logs, and the captured band from the selected protein preparation is consistent with the intended target.
The typical Kd value for a SOMAmer reagent is well in the sub-nanomolar range as determined by a solution binding method. We incubate the binding protein in with the SOMAmer reagent and determine the amount of complexes that are formed in a solution binding method where we capture the complexes through a tag or through a non-specific interaction. These are not determined by a method like surface plasmon resonance (SPR), so surface effects are not involved.

Can you make a SOMAmer reagent that will recognize a protein complex?

Yes. Selections have sometimes been performed against individual components or the complex and then tested against the complex. Please contact your Sales Representative for further information.

Study Design

How many samples are recommended per treatment group?

This depends on numerous variables: It is always recommended to consult a statistician on all aspects of study design. In the absence of such a consultation, we offer the following guidance:
The number of samples per group depends on the expected size of the physiological effect, the variation in the population to be tested, as well as other experimental parameters. The SOMAscan Assay is a discovery tool for hypothesis generation and is typically used to discover new biomarkers that can be further tested in follow-up validation studies.
In animal studies using genetically identical strains that are housed similarly, the variability is expected to be less than the variability within a group of humans. The median assay CV for the SOMAscan Assay is 5%. The variability of sampling genetically identical animals is approximately 3-5%. Combined, this results in an estimated CV of 8 – 10%. Suggested sample size estimates of animals per group (unpaired, with 8% – 10% CVs) as a function of effect size is shown in the table below.

Effect Size
(% change)
# Animals per Group
20% Very big
50% ~100
100% 20-30
150% 9-15
200% 5-8

If the samples are paired (drawn from the same animal over time or after treatment) the number of animals per group can be fewer. The number still depends on variables that require statistical analysis, but ideally, there should be a minimum of 8-10 animals per treatment group.

How many repeated measurements per individual do I need for a time course or dose response of cell treatment?

This depends on numerous variables: It is always recommended to consult a statistician on all aspects of study design. In the absence of such a consultation we offer the following guidance:
The number of repeated measurements per individual required for a study depends on the expected physiological effect size, the variation in the population to be tested, as well as other experimental parameters. The SOMAscan assay is a discovery tool for hypothesis generation and is typically used to discover new biomarkers. If the effect size is known to be large (> 2-fold), then fewer samples per group can be used.
For cell studies, our bioinformaticians recommend biological triplicates (n = 3) or greater, whenever possible. If there is a time course and the number of samples must be limited due to budget, our recommended strategy  is to reduce the number of samples at interior time points or doses to duplicates (n = 2), while retaining triplicate measurements (n = 3) for the time points/doses expected to vary the most from control. If a certain data point is critical for the experiment it should be run in triplicate or greater (n ≥ 3).  This is illustrated in the tables below:

Cell Line A
Dose Replica sub-total
Time 0 (vehicle) 10nM 100nM 1uM
0 3 3
24hrs 2 or 3 2 or 3 2 or 3 2 or 3 8 or 12
48hrs 3 3 3 3 12
Grand Total 23 – 27
Cell Line A
Dose Replica sub-total
Time 0 (vehicle) 10nM 100nM 1uM
0 3 3
12hrs 2 or 3 2 or 3 2 or 3 2 or 3 8 or 12
24hrs 2 or 3 2 or 3 2 or 3 2 or 3 8 or 12
48hrs 3 3 3 3 12
Grand Total 31 – 39

 

Do I need to include a media control sample?

It depends on the study question that is being asked. If you want to know what is secreted or consumed by the cells over time, then it is recommended to include a media control sample. If your question is primarily about the effect of treatment compared to a vehicle control, then it is unnecessary.

 

What are the best time points for my study?

It is recommended to use a previously known measure to estimate the best time points for your study. If nothing else is known, a cell viability measurement over a time course upon stimulation with drug can help to determine the answer.

Example data from previous studies:

Sample Type Dose Time Approximate # of analytes below FDR cutoff, p-value < 10-4.4.
(5%/10%)
# analytes with ³ log2 fold change below 10% FDR
Lysate 5uM 3 hr 15/30 8
Lysate 5uM 6 hr 32/57 9
Lysate 10uM 12 hr 13/23 8
Lysate
(four studies)
>1uM 24hrs 14, 16, 37 and 42 (5%)

 

How do I bridge previous sample runs to new runs?

Please contact SomaLogic to speak with one of our scientists..
Established calibration methods using endogenous pooled sample ‘calibrator’ references have been shown to reduce run-to-run variance on replicate samples. It has also been shown to improve consistency of run-to-run distributions of clinical samples provided the assumptions of data standardization (including randomization of clinical samples across plates) are met. We also run quality control samples (replicate samples from distinct populations) that are not used in data standardization and, therefore, may be used to evaluate post-calibration plate bias.
As time between runs increases, the opportunity increases for changes to occur in the assay that impact signal intensity and decrease the efficacy of standard calibration methods. Assay changes include necessary changes in consumable reagents (buffers, beads, control sample lots, SOMAmer reagent master mix) as well as assay protocol or instrumentation changes implemented to improve assay performance.
We strongly recommend designating a group of samples as bridging samples whenever one study is to be tied to another study.  These bridging samples should represent the independent groups or clinical populations that will be directly compared between studies. Including a set of bridging samples allows a direct comparison across the studies with a set of replicate clinical samples from the particular population/s of interest. Though this procedure mirrors our internal QC procedure, by using replicate clinical specimens a client will be in a better position to defend findings that arise from comparisons between independent sample groups assayed in different studies. In either case our QC samples provide an internal check that the inter-study calibration performed as expected.

Reproducibility and Interpretation of SOMAscan Results

What is the inter-assay variability between runs?

The SOMAscan assay has been well characterized for reproducibility with respect to inter-assay variation between runs. The SOMAscan assay has excellent reproducibility, of the 1310 SOMAmer reagents, the median total %CV is ~5% and 98% of the SOMAmer reagents have a median %CV of 15% or lower in plasma.

Does SomaLogic run disease-specific controls to evaluate disease sample/analyte reproducibility?

We cannot run these in the context of disease samples such as in plasma or serum because we traverse all diseases. However, we have evaluated the inter-run reproducibility of all proteins measured by the SOMAscan Assay where SOMAmer reagents have a median total %CV of ~5%.

How does SomaLogic report data?

We report differential protein regulation in RFU space as the SOMAscan Assay was designed for discovery purposes.

Are RFUs linearly proportional to the amount of protein present so that for a given protein, there is twice as much of that protein present in a sample with RFU 4000 vs. a sample with RFU 2000?

Relative florescence units (RFU) are a relative measure of protein quantity, which is a function of the unique characteristics (affinity, on-rate, off-rate) of each specific SOMAmer reagent as well as the assigned mix dilution. A two-fold change in signal may not be representative of a two-fold change in protein concentration.  However, it is appropriate to compare the RFU values for a specific analyte across all samples, which is the basis for most discovery analyses.

How is your data standardization controlled for quality?

A SOMAscan quality statement is delivered with each set of SOMAscan Assay results. Our data standardization procedure, which is designed to remove assay bias or batch effects, is composed of three main steps. The first step is hybridization normalization, which utilizes a set of control SOMAmer reagents added to each sample prior to hybridization in order to remove hybridization biases. The second step is median normalization and is designed to remove possible biases between samples due to variation in the assay, such as differences in liquid handling. The third step, “calibration,” removes inter-plate biases using a set of internal control replicate samples. The delivered data have undergone both of these normalization steps. Additional details and the acceptance criteria of each of these steps can be provided upon request.  Please contact us.

What was the distribution of the hybridization normalization scale factors and the distribution of the median normalization scale factors?

Hybridization and median normalization scale factors for each sample are included in the delivered ADAT data files. Hybridization scale factors are under the “HybControlNormScale” column heading and median normalization scale factors under the “NormScale_<dilution>” column heading in the ADAT file.

The relation between Dilution Linearity Range, Assay %CV, and F-Statistic (Mixed) is unclear; can you explain?

Each of these is an independent metric calculated to give some information about assay performance. Any one of these metrics, alone, may be sufficient for a scientist to gain confidence in the assay depending on the individual’s requirements.
Dilution Linearity Range (DLR) is a count of the number of points from a 16-point 2-fold titration series that fall within 25% of the expected value. The DLR may provide confidence because 1) this is the most straightforward metric of expected behavior when the protein is present and the SOMAmer reagent is capturing the protein and 2) linear behavior is ideal to make direct statements about protein concentration based on RFU readings.
%CV is the primary metric for assay reproducibility and this metric is very consistent across different matrices and species and is the only one of the three (%CV, DLR, F-stat) that is not strictly dependent on the population/samples used in the qualification experiment.  For this reason, some scientists rely on reproducibility metrics such as %CV alone and do not use the other two (DLR, F-stat) in their evaluation of the assay.
F-stat informs whether the small population tested has some variance above assay variance. Stated another way, F-stat informs us whether the SOMAmer reagent is ‘signaling’ in the population tested. The F-stat gives us information about the matrix/species rather than about the SOMAmer reagent performance but is useful to get a general idea of the expected number of proteins in the menu that may be measured differentially among a set of individuals. The population variance is established with a small population of healthy individuals and therefore not representative of the population in a clinical setting.
None of these metrics is definitive regarding the absolute performance of a SOMAmer reagent or of the assay. For example, DLR is based on unit linear behavior, which is not strictly necessary for an assay reagent to have excellent performance. Another particular issue is that the population selected for a future experiment may be very different from the samples/population used for the titration and F-statistic experiments. For these reasons, we combine these independent metrics into the Qualitative Score to make our most conservative statement about the matrix based on the experiments run.

Do you offer statistical support?

We can provide statistical support; the costs may depend on the study and analysis requirements. Please contact your SomaLogic Sales Representative for details.

Can we get the data in an Excel file or a file compatible with statistical analysis software?

The ADAT file is compatible with a variety of software tools. To open the file in Excel, do the following: File, Open, browse to the file location, select All Files (*.*) and double-click on the file. Follow the Text Import Wizard prompts or simply select Finish since the default settings conform to the ADAT file configuration.