DiNiro Documentation & Help

Left view :

1. Input Data:

The input data to DiNiro must be a Scanpy AnnData object in H5AD file format. The H5AD file contains your analyses, clustering, and plots created with Scanpy. When using DiNiro, you will have access to all the stored observations and variables in the uploaded AnnData object.
To use DiNiro you have two options:
1) On the [Upload Files] card upload your desired H5AD file. You can additionally upload your Transcription Factors (TFs) of interest as a line-by-line text file (DiNiro will only use these TFs during analysis).
Once you have selected the files to upload click on [UPLOAD] to save files and generate the single-cell data plot.
2) Alternatively, you can try out one of our two example data sets. Click on the button [GET STRATED] Below the data set you would like to try.

2. Plotting Choices:

The drop-down menu options are based on your pre-analysis, and they will appear as you name them during scRNA-seq data processing with Scanpy.
The single-cell plot is based on two main options: plotting-map and color-option:
1) Plotting-map is the dimensionality reduction method used for data visualization. You can use the [Type of plot] drop-down menu to choose among the already computed and stored plotting -maps in the Scanpy AnnData object. Typical methods are PCA, t-SNE, and UMAP.
2) Color-option is the coloring criteria to use for cell coloring such as cell type, tissue type, etc.
[UPDATE PLOT] button allows you to update the plot and try a different combination of plotting -map and Color-option.

3. Selection:

DiNiro performs differential co-expression analysis between two samples. Thus, selecting samples is a crucial step.
There are two types of selection methods to form your samples: CLUSTERS SELECTION and LASSO SELECTION:
1) CLUSTERS SELECTION is the straightforward method that allows you to form the samples from the single-cell clusters (a.k.a., cell types) available within your data. At least one cluster should be selected in order to form a sample, but multiple clusters can be also selected to form a sample. A cluster can be present in both samples. Use the [Select clusters to ADD to sample] drop-down menu to select the cluster and the [ADD] button to add it to the sample. Finally, use the [Type in an ID for sample] field to give your sample an ID (this will be useful later on when browsing the results). The same applies to making the second sample. When you click [CLEAR SELECTION] all the selections will be erased.
2) LASSO SELECTION comes in handy in case there are no clusters in the data in this case you can use this selection to manually add cells to each sample by drawing a circle around them. The first lasso selection will be automatically added to sample one and the selected cells will turn green, the text field will indicate the number of cells selected. The second selection will be added to sample two and the selected cell will turn red. Use [Type in an ID for sample] field to give your sample an ID (this will be useful later on when browsing the results).
[RESET SELECTION] allows you to erase the selection and start over.

4. Parameters:

For simplicity reasons, the parameter section is divided into two types of parameters: normal parameters and advanced parameters.

5. Computing:

This view contains the computation progress bar and the URL that you can use to retrieve your result later.

Middle view:

1. Single-Cell Data Plot:

This view is dedicated to interactively viewing the data. The user has control over the plotting map and the coloring criteria using the [Plotting Choices] view.

Right view:

This view contains a summary of the data such as the number of cells, number of genes, and plotting maps.

Results Interpretation :

DiNiro serves as an intuitive interactive platform for reconstructing and identifying transcriptional gene regulatory network modules that differentiate cell clusters. By comparing single-cell samples of interest, researchers can detect the gene that vanishes or switches regulation across samples. This facilitates exploring the results and focuses on modules that contain more relevant biological information. by looking at the gene modules that are enriched only in one of the groups the user can identify potential regulatory mechanisms that explain the difference across the compared groups. In all, this will assist in understanding the gene expression mechanism inside cells, and promote the research of disease pathology at the level of a single cell.

DiNro returns differentially co-expressed modules in the two input samples. A module is always formed from a TF and the differentially expressed genes regulated by this TF in both samples.
DiNiro offers three different views of the results allowing the user more flexibility when interacting with the results.

Table :

You can view the results as a table of five columns as follow:
Rank : We use a differential expression score to rank the outputted modules the results can be also limited based on this rank (see parameters).
Module TF: The TF that regulates all the genes within the module.
Genes : The module genes.
GO Terms: The top significant go terms are based on the enrichment analysis of the module genes.
Module Network: Module networks are displayed in this column. TFs are the blue triangles and genes are the orange circle the direct edges are code colors based on in which sample the interaction was found.

Network :

The network allows you to visualize all modules together and spot the genes that are regulated by more than one TF and also search for some specific genes of interest (useful in the case of large networks).
In top to left, you can type in or past the gene to search as comma-separated-values then click on [SEARCH] to start the search. If the genes are present in the network they will be colored in yellow.
The download allows you to take a screenshot of the current view of the network. The network legends are displayed in the top left corner.
In both the left and right bottom corner there are a bunch of buttons to interact with the network such as zoom in/out, etc.

Report :

This view contains a dynamic Venn diagram plot illustrating the distribution of the differentially expressed genes across the samples and also in form of a downloadable table.
Similar to the network view here you can interact with the result in different ways. You can get information on the distribution of the differentially expressed genes across the samples. You can also search for genes of interest in the Venn diagram plot using the search bar on top.