nCounter Knowledge Base: Assays & Applications

nCounter technology is ideal for a wide range of discovery and translational research applications, including gene expression analysis, solid tumor profiling, immune-oncology profiling, gene fusion analysis, single-cell gene expression analysis, miRNA expression analysis, copy number variation analysis, lncRNA expression analysis, and ChIP-String expression analysis. Additionally, multiple analytes can be profiled within a single experiment, allowing for maximum flexibility on projects where simultaneous digital detection of RNA, DNA, and protein is paramount.

There are three main differences between these two products:

Difference in Build: A Custom CodeSet has the gene-specific sequences built in to the reporter and capture probes whereas a TagSet would require the user to order gene-specific oligonucleotides.

Level of Multiplexing: A Custom CodeSet can multiplex up to 800 genes and a TagSet up to 192 targets.

Difference in Workflow: If ordering a Custom CodeSet, NanoString will provide all of the probes necessary for the reaction at the appropriate concentrations. For a Custom TagSet, NanoString provides the recommended sequence information, and you will need to order the oligonucleotides and prepare them at recommended concentrations before adding them to the hybridization reaction.

Because capture and reporter probes are added in excess in each nCounter reaction, a highly overexpressed target gene may prevent the detection of low-abundance target by saturating the available surface area on a cartridge. An over-abundance of one type of probe-target complex can reduce the chances that a low abundance target will be able to bind and be detected. In essence, a highly overexpressed probe will occupy more of the available binding “real estate” on a cartridge. Attenuation is a strategy to reduce the number of over-abundant probe-target complexes bound to a cartridge and therefore increase detection of less-abundant targets.

If attenuation is necessary, the first step is to run at least one sample to determine the total raw counts for all genes (no attenuation). In parallel, run the same sample with 90% attenuation by adding 180 pM of “cold” Reporter Probe oligo (Reporter Probe without the barcode) to the hybridization reaction containing 20 pM active Reporter Probe. If attenuation is necessary, 90% will be a robust attenuating factor for almost any gene. Minimally, this test requires only half a cartridge to process.

It is important that the un-attenuation sample counts be below the saturation threshold. If binding density is > 2, repeat the sample with ¼ of the RNA input. Reducing the input will ensure that the attenuation measurement will be accurate.

Unfortunately, tRNAs are not compatible with nCounter. The sequence composition, length, and distribution of sequence diversity in tRNAs are not amenable to our platform.

Yes. NanoString’s nCounter technology is based on a novel method of direct molecular barcoding and digital detection of target molecules through the use of color-coded probe pairs. The nCounter miRNA Sample Preparation Kit provides reagents for ligating unique oligonucleotide tags (miRtags) onto the 3’ end of target miRNAs, allowing short RNA targets to be detected by nCounter probes.

NanoString cannot guarantee the performance of a component if it has not been stored at the recommended temperature.  These reagents should be used at your own risk.

CodeSet: Improper storage of the CodeSet

Multiple F/T cycles can be detrimental to the CodeSet as they can damage RNA, DNA and fluorophore linkages; however, CodeSet performance is generally robust and may not be dramatically affected.  For example, if the -80°C fails, the CodeSet will likely maintain full functionality after brief storage at 4°C or even room temperature.  Customers have gotten robust data from Codeset that was at RT for 8 hours, however these results are not guaranteed.  If wanting to use a compromised codeset, it’s recommended to test with samples that have been run before on pristine codeset for comparison.

If the CodeSet is stored at -20°C, it may display some loss in signal after 1-2 months. This could introduce technical variance to the experiment thus, it is recommended to reorder a new CodeSet.  

Max/Flex Cartridge: Improper storage of the Max/Flex cartridge

If the Max/Flex Cartridge is briefly stored at -80C° (1 week or less), it can still be used without compromising its function. Prolonged storage at -80°C will damage the streptavidin binding surface, and a new cartridge should be ordered.

Prep plates: Improper storage of the Prep Plates

Prep Plates will freeze at -20°C or -80°C and should be reordered if frozen due to damage to the magnetic beads. Prep Plates that reach room temperature during transit should be replaced and not returned to 4°C.

Sprint Cartridge: Improper storage of the sprint cartridge

Buffers in the cartridge will freeze if stored at -80°C. Also, the valve membranes will freeze and lose the ability to open and close, resulting in lane blockages. The frozen cartridge should be discarded, and a new one should be ordered.

Generally, no. Bubbles will not affect the purification or the scanning of the cartridge.

Bubbles can sometimes appear in the cartridge – we believe it is related to atmospheric conditions in the lab. This is only an issue in the rare instance that an electrode ends up inside a bubble, in which case nothing will stretch in that particular lane. 

Sprint Reagent C  should be handled on its own as hazardous waste. This is our practice for disposing of our expired batch. However, a dilute mixture of Sprint Reagents A, B, and C should be fine for sewer discharge in the majority if not all jurisdictions. An extremely small concentration of sodium azide is not a hazardous waste concern, per se, but rather a concern of its tendency to form peroxide crystals on plumbing (which, of course, if shocked, can be explosive). However, this is easily mitigated by flushing copiously with fresh water after discharging sodium-azide-containing solutions.

In addition to the above, please refer to our SDS, and it is advisable to check their respective local sewer regulations.

Yes. The Sprint Reagent C should be slightly cloudy; that is the normal state of this reagent. This might appear as ubiquitous cloudiness or strands of opaqueness, especially near the bottom of the bottle. Clear and cloudy bottles are both normal and fully functional.

It is possible to purchase additional strip tubes and caps separate from the Master Kit.  These can be purchased directly from the manufacturer, BioPlastics/Cyclertest, Inc. (https://bioplastics.com/index.aspx), and the product information is as follows: Cap item #B56501 – EU thin-wall 12-cap strip, robust, indented flat, natural and Tube item #B56601 – EU 0.2ml thin-wall, 12-tube Strip, Extra Robust, Regular Profile, Frosted, Natural.

Reagents A & B are 500mL each. Reagent C is 125mL. (One bottle of each reagent is enough for 16 runs.)

Visible variation in the thickness of our cartridges is expected and will not impact assay results. In the extremely rare case that cartridge thickness prevents proper loading into the digital analyzer, please contact support@nanostring.com for further assistance.

The nCounter gene expression assay can use purified total RNA, raw cell lysates in guanidinium salts, blood lysate from PAXGene™, and amplified RNA directly added to the hybridization reaction. We have also obtained good results from total RNA isolated from FFPE samples, provided the RNA meets minimal integrity requirements (more than 50% greater than 200 nt).

The nCounter Analysis system is ideal for profiling DNA, RNA, and protein in any combination simultaneously. A wide range of sample types are compatible with most of our assays, including isolated RNA, cell lysate, FFPE, and fresh frozen tissue. Please check the product specifications or contact support@nanostring.com for more information on compatible sample types with your assay.

For most gene expression assays, we suggest 100 ng of total RNA or a lysate of 10,000 cells. More or less material may be used to boost signal or reduce sample requirements, but quantitation of extremely rare transcripts may be affected when less material is used in our system.

To a large extent, no. Differences in loading can be easily removed during CodeSet content normalization in nSolver. It is important to note, however, if the sample input is so low that the counts for your genes of interest drop below background, then normalization cannot compensate.

No, the use of a polyacrylamide carrier will not interfere with our assay.

If you are using purified RNA as your sample input, then 8 µl is the maximum amount recommended for an nCounter XT gene expression assay.

NanoString has determined that the purification methods for Probes A and B can be critical for optimal Elements assay performance; Probe B is best when purified by PAGE, while the purification method for Probe A is less important.

Two different purification methods are recommended because the assay is sensitive to probe cross-contamination. For example, if probe A is stored for any length of time with probe B, a very low rate of cross-linking occurs between probes even at very low temperatures. These cross-linked probes will result in higher background signal even in the absence of target RNA. Similarly, if the two probes are manufactured on the same production line (as might occur if they are both PAGE purified), then there is sufficient cross-contamination between probes to yield higher backgrounds. Thus, NanoString recommends using two different purification methods for Probes A and B to ensure that there will be no cross-contamination during manufacturing.

To ensure accurate sample temperature and to minimize evaporation risk, we advise using plastic consumables that are recommended by your thermal cycler manufacturer.

The goal of using a spike-in oligo with your miRNA panel is to introduce a known constant across all of your samples. In doing so, this internal control can account for differences in counts that would arise solely from small variations in purification efficiency from sample to sample.

All of our miRNA CodeSets are ready for you to add spike-ins, since the spike-in Reporter probes are automatically included in all CodeSets. If you do not add spike-ins, these reporter probes do not detect any target sequences and will appear as having no counts upon analysis. If you wish to add spike-ins, refer to the Gene List Excel file for your panel, and scroll to the bottom to see a list of spike-in sequences. Order the oligos that match the spike in Reporters.

When performing your spike-in experiment, add the spike-in oligos after lysis, but before purification and extraction. Please refer to our Tech Note for more details.

Yes. Lysates from primary isolated cells or cultured cell lines can be used in a nCounter GX hybridization without further RNA purification. The maximum sample input volume when using cell lysates depends on the type of lysis buffer used.

Nanostring has developed recommendations specific to primary isolated cells and tissue cultured cells.

More details are provided in protocol MAN-10051-03: Preparing RNA and Lysates from Fresh/Frozen Samples.

Yes. Tissue lysates can be used in a nCounter GX hybridization without further RNA purification. Best results are obtained if the lysis and homogenization is done in a concentrated guanidinium salt-based buffered solution. Many manufacturers’ RNA purification kits use a chaotropic guanidinium salt-based lysis step to break open cells and inactivate nucleases as a first step. Once completed, the lysate can be stored and used in a nCounter assay. If Buffer RLT is used for lysis, the maximum recommended input volume into the hybridization is 1.5µL. 

NanoString recommends a minimum of 5,000 (for SPRINT) to 10,000 (for MAX and FLEX) cell equivalents per nCounter XT hybridization reaction for most applications. The required number of cells for any given application will ultimately be dependent on the abundance of the mRNA targets of interest in the tissue sample to be assayed. As a guidance for the amount of RNA to be used in hybridization for a given sample type and codeset combination, refer to the binding density QC parameter in the nSolver User Manual. The binding density, reported in barcodes per square micron, responds almost linearly to the amount of sample RNA introduced in the hybridization within the optimal range specified for your instrument.

You do not always need to run technical replicates with gene expression assays; however, if you are a new user, it would be worthwhile to run technical replicates initially, as it will allow you to gain confidence in the technology. Biological replicates, however, should always be run whenever possible.

NanoString automatically includes probes against fourteen ERCC transcript sequences in every CodeSet. Six of these sequences are used as positive hybridization controls; the corresponding synthetic RNA targets are present in the CodeSet as well. Eight of the probes are used as negative controls, and the corresponding target transcripts are absent. Collectively, these internal controls allow you to determine the hybridization efficiency and non-specific background in your experiment. It is up to you to decide whether you require additional experimental or biological controls in your experiment.

It is unlikely, as capture and reporter probes are present in excess. However, the formulation itself may be compromised if the probes are diluted or if the total volume added to a hybridization reaction is altered from what the user manual recommends.

You may see a moderate increase in counts associated with longer hybridization periods. However, this difference normalizes away during data analysis.

Typically, the incubation conditions during sample hybridization and processing do not lead to DNA denaturation. Therefore, any DNA present in your sample should remain double-stranded and invisible to the probes, which need single-stranded targets to hybridize.

DNA contamination may interfere with your assay if it causes an overestimate of quantitation of the RNA concentration in the sample. To avoid this, follow the steps below when preparing your sample. Please contact support@nanostring.com for further assistance with sample preparation.

• Use an RNA quantification method that discriminates DNA from RNA, such as QuBit.
• For overestimations of several fold (as opposed to several log), careful data normalization may be able to remove bias from your dataset.
• If possible, increase the amount of RNA from the DNA-contaminated sample as long as total input remains below critical levels of saturation. If you observe that your DNA-contaminated samples lead to binding densities below 1.0, you can safely increase the sample input to boost the counts instead of relying on normalization alone. Contact support@nanostring.com for more assistance with determining your ideal input amount.

If you are using a pre-made CNV panel, we strongly recommend using AluI to fragment your DNA as we have verified that the AluI sequence is not present in any of the target sequences in the panel.

In addition, selecting an alternative restriction enzyme may impact your total number of counts, as the size of the fragments influences the binding kinetics of the probe to its target. Smaller fragments will exhibit faster hybridization kinetics and may result in more counts even if the number of molecules is the same, conversely, bigger fragments, as the ones you would obtain with EcoR1, may result in fewer counts.

If you are creating your own custom CNV panel, ensure that the restriction enzyme you select is not present in your panel target sequences. You may also wish to validate your experiment for the average fragment size you obtain with the restriction enzyme of your choice.

If you wish to concentrate your RNA before performing an nCounter assay, NanoString recommends using a commercially available RNA concentration kit. These typically work to both concentrate your sample and reduce impurities that could negatively affect your assay results.

NanoString cannot guarantee the performance of a CodeSet if it has not been stored at the recommended temperature. These reagents should be used at your own risk.

The nCounter gene expression assay can use purified total RNA from fresh frozen samples (25-100 ng), as well as purified total RNA from fragmented or FFPE samples (input will vary with fragmentation profile, see FFPE MAN-10050 and contact support@nanostring.com for more details). We have also obtained good results from amplified RNA from frozen or FFPE samples (1 or 10 ng pre-amplification, respectively, see low input MAN-10046), as well as from raw lysates of cell suspensions (lysis buffer and cell numbers will vary by cell type, see MAN-10051 and contact support@nanostring.com for more details).

Most NanoString assays require relatively modest total RNA concentrations (10-20 ng/µL for mRNA, 16.5-33 ng/µL for miRNA).  Nevertheless, although total yields for some samples may be sufficient, sometimes the concentration of the samples still does not meet these standard guidelines.  In other situations, the use of highly degraded RNA from (e.g.) fixed samples may require higher concentrations.  Thus, it may sometimes be useful to concentrate RNA further to obtain sufficient mass for use within the sample volume limitations of an nCounter assay. 

In general, there are three approaches one can use to concentrate RNA:  column concentration, precipitation, or evaporation.  There is not one universally ideal approach, as the best method will be dependent on the sample source and the assay being performed. 

The advantage of column concentration is that sample impurities like salts or organic solvents which may have been inadvertently carried over in the sample during the purification process will be removed.  Consequently, such columns are generally the preferred method of concentration for nCounter miRNA assays, as these are more sensitive to sample impurities (the miRNA assay has an enzyme-driven ligation step).   Columns for concentration can be purchased from one of several vendors.  Two columns which we recommend for miRNA assays (required for biofluid samples like serum or plasma) include one from Amicon (a size exclusion column for proteins but also works for miRNA: Millipore #UFC500396) and another from ZymoResearch (RNA Clean & Concentrator #R1015).  The disadvantage of using a column concentrator is that there will likely be some loss of total RNA mass.  Therefore, for some samples with limited volumes or limited initial RNA mass, there may be little improvement in final concentration. 

Precipitation of total RNA using a co-precipitator like linear acrylamide or glycogen can also be an effective method to concentrate RNA.  Recovery yields can be quite good even with low RNA quantities. However, the procedure is more labor-intensive, requires handling of toxic compounds, and may still result in some impurities being carried over into the final precipitate.  Therefore, this may not be ideal for nCounter miRNA assays run with biofluid samples where target concentrations are low and assay efficiency needs to be maximized. 

The use of a Speedvac is another option for concentrating RNA samples.  After evaporation, the sample can be re-suspended in the appropriate volume, though whatever impurities were in the original sample will also be concentrated within the final sample.  This method would therefore only be recommended for samples with high purity specs (260/280 and 260/230 ratios of 1.8 to 2.2), or for samples intended for use in any of the nCounter enzyme-free assays (that is, neither the miRNA assay nor any assay that includes a pre-amplification step).  

The goal of using a spike-in oligo with your miRNA panel is to introduce a known constant across all samples. This internal control can account for differences in counts that would arise solely from small variations in purification efficiency from sample to sample. Generally, spike-ins are not required if already using a robust normalization strategy such as the top 100 method as we recommend for samples consisting of tissues or cells.

All of our miRNA CodeSets are ready for you to add spike-ins, since the spike-in Reporter probes are automatically included in all CodeSets.  If you do not add spike-ins, these reporter probes simply do not detect any target sequences and will appear as having no counts upon analysis.  If you wish to add spike-ins, refer to this link, and under “Product Information”, select the Gene List Excel file for the appropriate panel, and scroll to the bottom to see a list of spike-in sequences.  Simply order the oligos that match the spike in Reporters.

When performing the spike-in experiment, add the spike in oligos after lysis, but before purification and extraction.  Please refer to our tech note here under “Support Documents” for more details.

Is this still okay to run?

• If the volume is low, but not gone: Yes
• If there is nothing left but thick residue or crust: No

If the volume is low, but not gone, then the evaporation was slow enough that the hybridization itself had time to occur. Before the sample is drawn out of the tubes, buffer is added by the Prep Station, and then the sample is used. This buffer addition will reinstate the necessary sample volume for successful purification.

If 50% or more of the sample is above roughly 200 nt fragment length, you can load slightly more RNA as a standard assay (150 instead of 100 ng total RNA) and get very similar raw counts.  Once that fraction drops below 50%, you’ll need to add increasing amounts of RNA to get equivalent counts to a fresh frozen sample.  There are ways to precisely calculate exactly how much one should add given a certain fragmentation level, but some educated guessing based on RNA fragmentation traces will also work: 

This approach depends on how concentrated the samples are-it may be necessary to concentrate the sample to achieve some of these higher input amounts and still be within the maximum sample input of 5µL. See Question 1 for details.

Typically, we recommend 100ng of input for FFPE-derived miRNA samples since the smaller targets from degraded samples generally have much less degradation than the larger rRNA and mRNAs. For extreme cases of degradation, one might need to increase input, but in most cases, it isn’t necessary. Most users can use the same input amount as fresh samples and then adjust if needed after the first cartridge.

The more common challenge for miRNAs and FFPE samples is slice thickness. Below ~ 8 µm, miRNA targets can leach out of samples during deparaffinization steps. Therefore, we usually recommend slices of at least 10 µm to prevent this leaching effect. 

Our low-input kit is designed for RNA applications and uses a targeted amplification approach with a low cycle number of amplification steps and primers designed to flank the region of probe binding. We currently do not have a protocol supporting a low input approach with the CNV assay which is a DNA application. Due to the varying efficiencies of primer binding and enzymatic activity, more subtle copy number amplifications or deletions may be difficult to accurately and reproducibly quantify.

NanoString has developed a proprietary assay design engine for probe design. The design engine contains algorithms that interrogate each target sequence in sequential 100 nucleotide windows, shifting along the target sequence one nucleotide at a time. The algorithm scores the 100-nucleotide sequences on a variety of sequence characteristics (hybridization efficiency, GC content, Tm, secondary structure, etc.) to identify target regions that fall within our ideal design parameters. Probe design rules include no more than 85% sequence homology between sequences in order for probes to be discriminatory and no more than 16 consecutive nucleotide matches.

The design engine screens all potential probe pairs against the target organism’s transcriptome to ensure specificity and by default we bias the final selected probe pair towards a region common to all transcript variants for each target gene whenever possible. In some cases, it is possible to design probe pairs that distinguish different transcript variants or target specific regions of a gene although this is dependent upon the specific sequences within each region.

The target sequences and associated probe pair data are included in a CodeSet Design Report that is sent to the customer for review and approval prior to the start of the manufacturing process.

The miRNAs included in the nCounter miRNA Expression Assays have been curated to ensure that only biologically significant miRNAs are included in the panel; the nCounter Human v3 miRNA panel contains a probe for all miRNAs that are denoted in miRBase 21 as “high confidence”. In addition, NanoString applies a set of proprietary metrics such as observed read ratios and expression analytics to screen potential targets prior to inclusion in the panel. NanoString also performs a scan of the current literature to ensure that only actionable and clinically relevant miRNAs are included in the miRNA panels. Altogether, each miRNA panel contains a comprehensive set of miRNAs that are biologically significant and ideal for targeted discovery and validation experiments.

Working with xenograft RNA is a particular strength of NanoString technology. We have extensive experience with multi-species designs, particularly the mouse-human xenograft tumor model. Designing a CodeSet that targets mRNAs from each species is relatively straightforward; we simply check all transcriptomes likely to be present in the reaction for cross-hybridization. In addition, our nSolver software allows you to create custom annotations for each targets, so it is easy to assign them to different pathways or species for downstream analysis.

Yes. By default, NanoString will design your probes to recognize as many isoforms of the gene as possible. However, if you would like to identify splice variants, it is possible to design multiple probes for one gene. Each splice variant will count as one “gene” in your final gene list. Please contact bioinformatics@nanostring.com for more information.

In all of our probe designs across organisms, we design to what is considered the reference sequence by NCBI unless otherwise specified in a custom project. The NCBI mouse reference sequence is from C57BL/6J mice. Our probes are robust to small changes in the actual target sequence, which makes them insensitive to most variation from the reference sequence that exists between strains. To determine how well a probe will work against a variant target, we compare the percent identity of the original, targeted sequence to the variant (using BLASTn). Any target with a percent identity of 95% or greater is likely to be targeted at similar efficiency to the intended sequence. Thus, in order for probe efficiency to be significantly altered, there would have to be more than 5 bases that differ between the reference and the alternate strain.

The miRNA assay is designed to digitally quantify mature miRNA. It cannot detect primary miRNA (pri-miRNA) or precursor miRNA (pre-miRNA). The assay includes a step to ligate an oligo (called a miRtag) to the 3’ end of mature miRNA molecules in order to provide the capture and reporter probes a sufficiently long molecule for hybridization. This lengthening with the miRtag is required because the NanoString probes hybridize to approximately 100 consecutive nucleotides and mature miRNA is typically 22 nucleotides in length. In addition, the double-stranded nature of primary and precursor miRNA and their inaccessible 3’ ends interfere with the ligation reaction. The ability to specifically detect only mature miRNA allows direct quantification of the molecules directly involved in gene regulation. For more information on the miRNA assay, please refer to the miRNA Assay Manual.

NanoString chemistry is based on hybridizing anti-sense capture and detection probes to a 100bp segment of the target of interest. Fragment size of the nucleic acid sample can modulate the efficiency of this reaction under certain conditions.

nCounter technology requires that RNA or DNA is sufficiently fragmented to allow for hybridization of the capture and detection probes. For optimal performance, we recommend that at least 50% of the sample be fragments 200 nt or larger, as determined by Agilent Bioanalyzer®. Total nucleic acid input may be increased according to fragment size distribution, per our Tech Note, if the sample is highly degraded.

The optimal upper limit of nucleic acid fragment size is 800 bp, and counts will decrease linearly as the percentage of fragments above 800 bp increases. For mRNA and lncRNA targets, these are susceptible to nicking during hybridization reaction, and ultimately most fragments will be reduced in size so that the percentage of fragments over 800 bp is negligible and will not interfere with the efficiency of the assay. DNA targets, however, are stable at 65°C, and will not reduce in size throughout the hybridization reaction. It is therefore critical that DNA samples be thoroughly fragmented by AluI digestion or sonication.

Yes. It is possible to design probes to distinguish between precursor mRNA and mature mRNA sequences. Introns and exon/intron junctions can be used to detect precursors, whereas exon/exon junctions can be used to detect spliced mRNAs. Simply tell us what you would like to detect and we will design the necessary probes.

NanoString probes are designed to span exon-exon boundaries so that only mature mRNA species are counted. If you wish to detect pre-spliced transcripts, simply submit a request to bioinformatics@nanostring.com and we can design the correct probes to fit your needs.

Yes, internal reference and housekeeping genes are used synonymously in our literature. Internal reference genes are a subset of genes within your CodeSet that have low variability across sample types and high counts, regardless of their function. Usually these genes play generic, non-cell specific roles, such as in metabolism.

The Ligation Positive and Ligation Negative control sequences are pieces of ERCC (External RNA Control Consortium) sequences that were trimmed to have profiles similar to natural miRNAs. These sequences were tuned and refined by NanoString to perform robustly with our miRNA assay.