The table displays the probe annotations that are in use for the current analysis.
To customize or edit these annotations, click the Download CSV button and save the CSV file locally. This file can be edited and then imported back into the advanced analysis. See the comments at the top of the CSV file for more details regarding this process.
Use the Import CSV button to import probe annotations for your analysis. This can be either a probe annotation file that you've obtained from NanoString or a file that was previously downloaded and customized.
A quick analysis will perform experimental setup QC and differential expression testing using a single sample annotation field to group samples. Note that the baseline reference for the selected sample annotation is specified on the sample annotation page. Warning: If normalized data was chosen, do NOT run Quick Analysis. Only Custom Analysis should be used for Normalized Data.
This option allows you to choose which modules to run and customize all of the settings. Warning: If starting from Normalized Data as input, "Normalization" module must be deselected.
Select the type of experiment that is being analyzed. After inspecting the current data, nSolver has attempted to choose the correct value for this setting. You can change this if necessary.
A standard experiment with data from a single RLF.
MultiRLF Merge (standard experiments merged)
Selection of this option generates descriptive plots, including unsupervised clustering heatmaps, principal component analysis (PCA) plots and summaries of covariate (annotation) distribution across samples. This option is recommended for all analyses. These descriptive plots can be broken down across subsets of probes by selecting an appropriate probe annotation.
Selection of these options specifies that the advanced analysis module should attempt to normalize the data. Normalization options vary by analyte type. Normalization is recommended.
For each gene, The Optimal method infers differential expression with respect to specified covariate(s) using a negative binomial mixture model for low expression probes or a simplified negative binomial model for high expression probes. The Fast method uses the simplified negative binomial model for all probes. In situations of the mixture model algorithm not converging, the simplified model will be used instead. If the simplified model does not converge, the loglinear regression model will be used. High or low expression is determined by how high the probe mean is across all samples relative to the negative controls. At least one covariate must be selected as the predictor. Analysis will take into account the selected confounders but results will only be displayed for covariates designated as predictors.
This module condenses each gene set's expression data into a pathway score, creating a more manageable dataset of samples X pathway scores. The module allows for exploration of scores relative to selected covariates of interest (e.g. change in MAPK pathway scores relative to treatment type). Currently scores are based on the 1st principle component of the pathway probes in the dataset. At least three samples need to be present for this module to work.
Estimates cell abundance based on expression of cell-specific marker genes. Each cell type's abundance is estimated with the mean of its marker genes' expression. A variety of summary statistics and plots are created, including QC plots verifying the utility of these marker genes in any particular dataset.
The module provides descriptive plots for up to 15 user-selected probes relative to user-defined covariates of interest. Depending on whether a single probe or multiple probes are selected, bivariate or multivariate analyses plots are generated. Furthermore, the module provides trend plots when trending covariates are provided.
This module enables descriptive analyses of groups of probes corresponding to the same gene or genomic locus. These probes may target the same analyte types (e.g. ERK protein and phospho-ERK protein) or different analyte types (CD14 mRNA and CD14 protein). Use this module to investigate questions like how a gene and its protein are co-regulated and how they jointly respond to biological variables.
This module requires values in the Cell.Type column of the probe annotations.
This module requires a multi-analyte dataset and values in the Related.Probes column of the probe annotations.
Removes probes from the analysis based on a specified threshold count value and observation frequency across all samples. Probes that fall below the threshold at a frequency greater than the specified observation frequency will be removed from the analysis. To change our defaults, first de-select the "auto" checkbox.
Threshold count value: (min = 0, max = 100)
Observation frequency: (min = 0, max = 1)
Specify Fusion Parameters: adjust the p-value threshold for Junction probe detection and End probe imbalance expression.
Reassign SNV References by covariate type (default "Is Reference" refers to the assignment made during nSolver experiment creation) or by file name (manual selection). Adjust parameters defining the detection thresholds (Detected and Not Detected), based on the relative ratio of variant probes to reference probes (log2FC) and p-value. EM (Expectation maximization) performs an iterative refinement of background estimate after an initial call. Debias performs centering of reference probe data to prevent systematic bias due to normalization. By default, these options are selected; they work optimally in most circumstances. If you are troubleshooting unexpected results, uncheck EM and Debias one at a time and review the results or adjust the thresholds, as appropriate.
Each pathway score is derived using the 1st principal component of a gene set's data, but there is no guarantee a gene set's first PC captures interesting biology.
Plot Pathway scores vs :Each gene's expression data is regressed on the variables selected here, and pathway scores are fit to the residuals of these regressions.
Adjust Pathway Score for: If you think an uninteresting variable or a technical variable could be highly correlated with gene expression and therefore threatens to hijack the first PC, select that variable as confounder to scrub its influence from the data before calculating pathway scores.
Select sample annotations by which samples shall be grouped for analysis and display.
Specifies the gene annotation column that identifies cell type specific genes.
By default, cell type specific probes that do not mirror the expression of their counterparts will be removed before estimation of cell type abundance. Choosing the "Use All Probes" options skips this QC measure and uses all of a cell type's probes to measure its abundance.
Defines the significance threshold for reporting a cell type abundance estimate. Cell types whose evidence for cell type-specific expression does not meet this level of evidence will be discarded. By default, this value is set to 1, returning results for all cell types regardless of how well their probes exhibit cell type-specific expression in your data. This choice relies on the accuracy of the literature around these probes (Bindea 2013). Choose a value of 0.05 or lower to only see results for cell types whose quantification is further supported by your data.
Use these options to choose how expression activity abundance estimates are displayed. Raw cell type abundances show the estimated abundance for each individual cell type, e.g. T-cells, NK cells, etc... It is often the case that all the raw cell types scores are highly correlated due to variable amounts of total immune infiltrate in your samples. When this occurs, it can be useful to look at relative cell type measurement to reveal subtler patterns in the TIL landscape.
Relative cell types show contrasts between raw cell types, e.g. CD8 T cells vs. Tregs or NK cells vs. total TILs.
You must select at least one of raw and relative cell type abundances.
Running the Optimal model is computationally intensive and run time is proportionate to data size and number low expression probes. It may take several 10s of minutes depending on the data size and count distribution.
All genes are fit with the negative binomial model.
GSA requires that a valid probe set annotation column be selected on the General Options tab.
Summarizes the statistical significance of all genes in a pathway using their mean squared t-statistic.
GSA will utilize the value of the Probe Sets Column selected on the General tab.
This module requires values in the KEGG.Pathways column of the probe annotations.
| KEGG ID | KEGG Pathway |
|---|
Enter up to 15 probes for further analysis. If five or more probes are entered, PCA plots will also be generated. Probes used as housekeepers or removed from the analysis via a low count threshold will not be included in the output.
Expression data will be grouped by the levels of the selected annotations. At least one annotation must be selected.
Creates trend plot for a data series defined by the Series ID across a set of levels identified by Interval ID. The interval ID can be an ordered categorical or continuous variable. Additionally, trends across distinct sample annotation groups can be examined by specifying an optional stratifying annotation.
This option generates a network that best describes the conditional inter-relation between your selected probes. You can opt to adjust for a covariate that is expected to influence these probes.
For each set of related probes, this function creates trend plot for a data series defined by the Series ID across a set of levels identified by Interval ID. The interval ID can be an ordered categorical or continuous variable. Additionally, trends across distinct sample annotation groups can be examined by specifying an optional Stratifying Variable.
In order for many of the analyses performed by this module to provide robust results, multiple samples are needed. A minimum of 3 samples per condition is recommended, i.e. at least 3 samples per analysis involving continuous variables and at least 3 samples per level for categorical variables.
| Categories |
|---|
Warning: A non-default cell type gene annotation column has been selected, which may interfere with calculation of relative cell type abundance.
Calculation of relative cell abundance relies on a .csv file specifying a "cell type contrasts matrix". The default cell type contrasts matrix assumes that the default cell type gene annotation column has been used. Use of a custom cell type gene annotation column may limit the utility of the default cell type contrasts matrix: new cell type names will not yield relative cell type abundance results, and removed cell type names will eliminate their associated cell type contrasts.
The default cell type contrasts matrix uses the following raw cell type measurements:
To calculate the following contrasts:
Warning:
The nSolver Advanced Analysis doesn't support CNV or miRNA data.