# OmniBrowser GuideBook

# 1. Intruduction

OmniBrowser is a powerful web-based platform, which focuses on single-cell data exploration. Here you have the authority to access to the continuously updated comprehensive single-cell database. It provides flexible data retrieval system, interactive data visualization, and cutting-edge analysis. What is more, all these functions are command-line free! The software allows scientific and clinical users, even ones without programming experience, to quickly investigate massive amounts of single-cell data.

Welcome to the Home Page of OmniBrowser, if you are new users, please click on the Get Started or Start your free Trial button to begin with registering a 15 days free trial account. The trial account can access and unlock certain number of datasets for free. To access more datasets, a purchase can be made following the introduction on the page, or contact our BD department for more information.

If you have already registered or have an account, please click on the login button to enter the OmniBrowser.

After login, you are on the Dashboard Page. There is a navigation bar on the left side of the page. From top to bottom, the buttons are: Dashboard, Search, Search Gene, Normal Cell Atlas Page, Treatment Page, Plot Theme, Save Report, Update List, Help, FAQ, Feedback and Account infomation in turn.

Update List: Click on Update List icon to show data and function update list. The link in study column will bring you to the Load Dataset page of the corresponding dataset.
Help: Click on Help icon to check the OmniBrowser GuideBook.
FAQ: Click on FAQ icon to see requently asked questions and answers.
Feedback: Click on Feedback icon to send your feedback on any aspect of OmniBrowser to us.
Logout: When your mouse hovers over Logout button, a window will pop up to show the account infomation. User can reset password or logout of the current account in this window.

# 1.2 Dashboard

On Dashboard Page, the statistics of curated data are visualized according to different species and tissues. On the left area, the different species and tissue can be selected, once the icon is clicked on, the corresponding data statistics will be viewed on the right area. the filtered datasets can be checked on the Search Page.

# 2. Topic Page

# 2.1 Treatment

The Treatment Topic Page summarizes all tumor treatment-related datasets in the database with manual annotation. Seven major Treatment Categories, such as small molecular inhibitors, monoclonal antibodies, and biological response modifiers, and the statistics of treatment objects, tissues, library preparation methods, and disease ontology involved in each category are filters for users to quickly locate interested datasets by clicking.

Click on one icon of Treatment Categories to view the statistics of the selected datasets, showing by pie charts of Treatment Object, Tissue and Library Preparation Method, and histograms of Disease and Cell Type. Press Ctrl and click on multiple icons of Treatment Categories to view the statistics of multiple categories together. Click on Clear All to remove all the current filters.

Click on Show Study List to view the filtered dataset list. The treatment category of each dataset will be marked in Treatment Categories column. Sometimes, a single dataset may cover multiple treatment categories. Multiple treatment categories will be shown when hovering on the icon. Click on Hide Study List to hide the list.

More filters, such as Disease Ontology Name and Treatment Applied in Study, can be chosen to track down the datasets of interest more accurately. When the terms in Disease Ontology Name are selected, related terms of Treatment Applied in Study will be shown in white background; When the terms of Treatment Applied in Study are selected, related terms of Disease Ontology Name will be shown in white background. Both related terms will continuously increase as cutting-edge studies are included in Omni Single Cell Database.

Click on Search Gene on the top right corner of dataset list will jump to the page of Search Gene. The filters selected on Treatment Topic Page will be inherited.

Select datasets by the buttons in Operation column and click on Cross Analysis to show selected datasets. Then, click on Cross analysis button on the top right corner to jump to the page of Cross Analysis.

# 2.2 Normal Cell Atlas

The Normal Cell Atlas Topic Page displays the single-cell studies of normal tissues included in the Omni Single Cell Database. The normal cell data covers 70 human tissues and 50 mouse tissues. Click on the tab of Homo Sapiens or Mus musculus to switch the species. Click on the drop-down list of Tissue Search and choose one tissue to view the cell type distribution in multiple-tissue atlases and single-tissue studies.

Click on Show Study List to show all human/mouse datasets including normal cell data. Click on Hide Study List to hide the list. In the panel of study list, click on Search Gene on the top right corner of dataset list will jump to the page of Search Gene. Select datasets by the buttons in Operation column and click on Cross Analysis to show selected datasets. Then, click on Cross analysis button on the top right corner to jump to the page of Cross Analysis.

In the left area, the proportion of cells of each tissue in the multi-tissue study is displayed in a pie chart, and the cell types of all tissues are shown in a histogram as default. Click on one tissue in the pie chart will show the cell type distribution of this tissue in histogram, and click again to unselect the tissue. In the right area, the type and number of cells in a single tissue are displayed in a histogram.

The multi-tissue cell atlases support Comparative Analysis function. Click on Comparative Analysis to jump to the page of Comparative Analysis. Species and tissues can be selected on the left. Cell type distribution of selected tissue among different studies will be shown in histogram(s) in the middle. Specific cell type can be searched and displayed by setting the parameters on the right.

Click on the tab of Homo Sapiens or Mus musculus to switch the species. Tissue can be sorted by Alphabet or Occurrence in multi-tissue atlases. Click on the drop-down list and choose one tissue or click on the tissue beside the heatmap directly to view the cell type distribution in multiple-tissue atlases. The subject of each multi-tissue atlas can be shown by hovering on the study ID. The number displayed when the mouse hovers over the color block represents the percentage of the number of cells in the tissue to the total number of cells in the study. Tissues not involved in the study are shown in grey. The tissue heatmap can be downloaded or added to report by the toolbar on the top right corner.

Cell type distribution of selected tissue among different studies will be shown in histogram(s) in the middle. The subject of each histogram is shown as the title. The number displayed when the mouse hovers over the color block represents the percentage of the number of this cell type to the total cells in the tissue of each study separately. The cell type histogram(s) can be downloaded or added to report by the toolbar on the top right corner.

Specific cell type can be searched and displayed by setting the parameters on the right. Click on Show All to view all the cell types in the selected tissue among different studies. If there are too many cell types to display, click on Show Top and input a total number of displayed cell types.Click on the drop-down list and choose one cell type or click on the icon of one cell type directly to highlight this cell type and the percentage will be shown in the histogram(s). Click the icon again to unselect the cell type.

# 3. Search

OmniBrowser provides multi-dimension search strategy to retrieve the dataset of your interest efficiently. There are summary function and search function. The summary function visualizes the aggregated data information of the whole database on Dashboard Page. For the search function, there are common search and advanced search. Common search provides different parameters to trace down datasets more rapidly; advanced search supports more complicated AND/OR logical operation and more detailed filters to retrieve the dataset more precisely.

# 3.1 Study(Dataset) Search

In the Search Page, the upper area is the search function area, you can switch between Common Search and Advanced Search tabs. Below the search function area is the result area.

# 3.1.1 Common Search

Input keywords in the search box for title or abstract as the filter to retrieve studies. There are six drop-down lists, where various parameters can be chosen as filters to track down the datasets of interest. There is a search rule notification on the right as guidance, which displays when the mouse hovers on the icon. It illustrates the search rules’ difference between the category level filters and intra-category level filters. Apply button needs to be clicked after selecting in each drop-down list, and Clear All button is used to remove all the chosen parameters.

# 3.1.2 Advanced Search

Advanced search provides a flexible search strategy to get more detailed and customized search results. There are two sub functions can be used in this tab:

Add Filter Rule: This function is used to add specific items of common search filters, which means the filter rule can be more detailed to track down the datasets of interest. Detail items of one common search filters can be selected at one search and the search logic between the items is OR, which means datasets containing either item will be shown in the result.

Add Join Group: AND/OR/NOT AND/NOT OR logic operation can be flexibly applied in this function. There are multiple common filters and items can be selected, and based on the logic operation, complicated search filter combination can be created and applied to track down datasets more accurately.

After inputting the select filters, click on Apply Filters button can be used to search or Clear All to remove all the selected parameters.

# 3.1.3 Display Filters

There are two authority control buttons above the dataset list. The Permission button is used to filter the datasets which can be viewed or not, or can be downloaded or not. The Dataset button is used to filter the datasets which are uploaded by user or from Omni Single Cell Database.

# 3.1.4 Dataset List

Datasets are listed according to studies. The overview of the study shows the PubmedID, journal, Publication date, author and abstract. For the datasets, there are also preview of the details of the dataset, like the ID, species, description, research topic and so on. The tags or pie charts beside each dataset ID give a quick view of the relationship between datasets in a study.

Dataset tagged with ALL means that this dataset contains all the single-cell data of this study. Dataset tagged with UNIQUE means that the single-cell data of this dataset are exclusive. Dataset with pie chart means that there are overlapping cells (with other datasets) existing in this dataset. When hovering on the pie chart, a detailed chart will pop up to show the statistics of the overlapping cells between this dataset and other datasets in this study.

Click on the link of dataset ID to jump to the Load Dataset Page of a single dataset. In the right end of the dataset is the Operation bar, there are buttons in this area:

click on the button beside Operation to select all datasets belonging to the study and add these datasets to Cross Analysis dataset list;
redirect to the corresponding single-data analysis page;
add one dataset to **Cross Analysis **dataset list at a time; For trial accounts, this button could be locked. Click on either of the buttons will lead to a locked page. Click on Purchase to unlock this dataset if you have the authority, or please contact our BD team for more information.

# 3.2 Gene Search

OmniBrowser enables gene search among large-scale published single-cell data by Search Gene function. Click on the icon of Search Gene on the navigation bar, or click on Search Gene button on Search page to jump to the Search Gene page. The latter operation inherits the results from Search Page, if you need to search among all datasets, please click on Clear all button before searching gene.

To search one specific gene, input and then select your target gene in the drop-down list. Click on the Related Genes on the right of the input box to show the ideograms with interacting genes and paralogous genes. The External Data tab shows the distribution of pathogenic-related genes recommended by ACMG, which can provide more references for users to do gene research.Decide which annotation to be applied in the search by choosing the annotation label type on the top of the parameter panel, and the cell clusters under the gene input box will change correspondingly. Selectable annotation labels are as below:

Cluster Name includes the annotation given by the author of the original article, which varies greatly among different studies. Datasets with cell type annotation given by the author will be categorized as Annotated, choose the filter in the drop-down list of clusters to filter the clusters to be searched.
Marker Gene Anno includes the manual annotation given by Abio data curation team. The manual annotation is done for datasets without annotation given by the author according to the dimensional reduction figures in the article. Choose the items in the drop-down list of clusters filter the clusters to be searched.
Cell Ontology Name and CellO Classification include the annotations automatically calculated according to Cell Ontology and CellO classification separately.
Scibet Major Type, Scibet Sub Type, Scibet Tabula Muris, and Scibet Tabula Sapiens provide predicted annotations by Scibet algorithm according to different references.

Besides positive search, negative search is also available by switching IN to NOT. When NOT is chosen and applied, datasets in which the input gene is not expressed will be shown in the result.

The result by F-test with the p-value smaller than the input value, or by gene expression satisfying the input range.Input the min or max value in the boxes of ln(TPM+1) to achieve quantitative search. F-test can be filtered by inputting a value in the input box of F-test.

Click on Add Gene to add a gene to be searched together. Click on AND or OR to select the search logic.

Click on Yes to remove the batch effects of the gene expresion level among datasets presented, so they can be compared at the same level.

After all the search conditions are set, click on Search to search, and the result will be shown in List tab and Summary tab.

In List tab, the total number of studies and datasets satisfying the filters will be shown on the top right corner. The result can be sorted by expression or F-test, click the drop-down list on the top left corner and set the parameters. Gene expression among all clusters in this dataset will be shown in box plot for single gene, and bubble plot for multiple genes.

In Summary tab, gene expression of each cell cluster among different datasets will be shown in a bubble plot. The result can be sorted by expression or F-test, click the drop-down list on the top left corner and set the parameters. 20 datasets and 20 clusters will be shown as default. Choose datasets or clusters of interest to be displayed in the corresponding drop-down list. The plot can be exported by clicking on Save As Image button on the top right corner of the bubble plot.

# 3.2.1 Cell×Gene Expression Matrix Download Service (Cross-dataset)

Gene expression on specific cells among multiple datasets can be downloaded as matrices with selected cell meta information for downstream analysis. The gene expression downloaded is the TPM or normalized expression of each cell, which will not be affected by batch correction option. This function can be unlocked by purchasing Data Quota and is now available on both Search Gene Page and Dimensional Reduction Page of each single dataset.

Data Quota of each cell×gene expression matrix to be downloaded = Number of Selected Genes × Total Cell Count.

All the cell meta of each dataset will be downloaded together with expression matrix and will not cost any Data Quota.

Downloaded matrices can be viewed in Download Expression Matrix List and can be repeatedly downloaded without extra deduction of Data Quota.

Download history is shown in Download Expression Matrix Page, on which the download of expression matrix of single dataset will also be shown in the list. The detailed information of each task can be shown when hovering on the corresponding item. Used Data Quota and Total Data Quota are shown on the top-right corner of the list, and can also be found in account information by hovering on the icon in the navigation bar.

To purchase Data Quota for expression matrix download, click on the Purchase Data Quota button in the bottom of the navigation bar to jump to Omni Product Store and complete the payment. Don't close the page of OmniBrowser during payment according to the instruction. When the payment is completed successfully, click on Purchase Completed button to refresh the Data Quota.

# 4. Single-Dataset Analysis

OmniBrowser inherits the analysis results from the author. If it is not available, a unified processing flow is applied to generate a complete dataset, and there will be a real-time visual interactive interface to display the dataset. In this processing flow, expression matrices and sample annotations were collected from GEO, GSM and other sources. Then after a standard curation process and three quality assurance checkpoints by different specialists, a structured data object was generated. The structured dataset contains matrices, study metadata, cell and gene annotations, tSNE/UMAP coordinates, marker genes and so on.

The analyses are based on the available cells in the dataset. For most of the functions, the processing time for the analysis or the calculation will take less than half a minute.

# 4.1 Load Dataset

When entering a dataset, the default page is Load Dataset Page. In this page, there are three parts: Introduction, Similar Cell Ontology and Marker Gene Ontology.

Introduction: in the introduction tab, the number of cells and other tags of a paper are shown, following by author and abstract. Article outline figure can be viewed via the View Figure button. View Source button will redirect you to the publication. You can also use keyword to search other studies by clicking on the title of this publication.

A Note was added to explain the source of cluster name, tSNE, UMAP coordinates and whether the dimensional reduction plots are satisfactory. Disease Ontology Name and Disease Ontology ID are also added in the note if any. When the research is about cancer, there will be additional corresponding description. For datasets containing treatment information, a description will be displayed under Note.

Similar Cell Ontology: A table of datasets that have the same or similar cell type with this publication is listed. You can visit the listed datasets by clicking on their titles. The Distance parameter can be adjusted to show relevant datasets with different ranges under each specific cell cluster tab, 0 means the same cell type, larger number represents a larger cell ontology distance. Default is 1, which means the listed datasets include cell ontology distance less or equal to 1.
Marker Gene Ontology: The GO entities of for each cluster are listed here. View Detail List will lead you to the Gene Set Analysis Page to show more details of GO analysis. NS represents Name Space, including three types: molecular function(MF), cellular component(CC) and biological process(BP). P_fdr_bh means false discovery rate (Benjamini–Hochberg procedure), also named Q-value.

# 4.2 Dimensional Reduction

Dimensional Reduction is a fundamental but powerful function provided by OmniBrwoser. For visualization and further analysis of single-cell RNA sequencing data, PCA was used to reduce dimensions, then tSNE and UMAP coordinates were calculated. After clustering, OmniBrowser offers interactive 2D and 3D views of tSNE and UMAP results. Various parameters are available to adjust the visualization of the dimensional reduction plot.

# 4.2.1 Plot Module

Projection: You can visualize the dimensional reduction results in 2D or 3D view on Dimensional Reduction page. Methods can be toggled between tSNE and UMAP on the right-side panel. You can scroll Mouse Wheel to zoom in/out or press and hold the Right Mouse button to rotate in 3D view.
Scatter Size: adjust the sliding node to set the pixel of points in the plot.
Scatter Opacity：adjust the sliding node to set the opacity of points in the plot.

# 4.2.2 Category Module

Category View will show up by default on Dimensional Reduction Page. Plot can be colored by different groups of annotation labels, e.g. Cluster Name, Cell ontology name.

Clicking the triangle symbol on the left end to view the detail of each annotation label. In the right side of the annotation label, there are regroup button to combine different sets of annotation into one annotation label, clone button to replicate the annotation, save as image button, add to report button and view button. When clicking the cell group name beneath the annotation label name, the selected cell group will be highlighted on the dimensional reduction plot, and a dashed box on the cell group name. Unclick the bar can remove the highlight. Also, the color of each group is auto-generated to ensure that the color contrast of adjacent points in the chart is big enough for a good presentation.

Regroup function is both available for original annotation and customized annotation. Select cell clusters to be regrouped, and then click on Regroup button, a window of regrouping parameters will pop up.

Choose another annotation label to be combined with the current selected cell clusters. The clusters beneath will be changed correspondingly. Selected the clusters to be combined, and decide whether to maintain the annotation for cells unselected. Choose 'yes' will assign all the unselected cells to undefined, and choose 'no' will maintain them. Undefined cells will not be brought into downstream analysis. Then name the new category, and the two sets of annotation will be combined and assigned to the new cell clusters. The new category will be automatically generated as a customized annotation label, whose font is in blue. The new category can be directly viewed in Differential Expression and Correlation page, but calculation needs to be done for Marker Genes and Metadata page. Click on Downstream Analysis button to run the calculation and wait until the complete notification pop up on the page. For functions available for a new category, please refer to 4.2.3.

# 4.2.3 Customized Annotation

There is an import button on the upper right of the category module, which supports importing json format customized annotation file, and the sample file can be viewed when the mouse hovers over it.

After clicking on the clone button, there will be a pop-up menu to input the new set of cell annotation label name. The replicated cell annotation label name is blue; Besides the cloned annotation label, there are various buttons here:

Recluster: select at least one cell group, clicking on the Recluster Selected button, and there are two parameters to be set: Resolution - the higher value of the resolution, the larger reclustered cell group count; Cutoff - used for setting the low limit of cell group count value, default value is 0.5%. If the cell group count after re-clustering is smaller than the cutoff value, this cell group will distributed to the original cell group.

Merge: select at least two cell groups. Click on Merge selected button and rename the new cell group name to merge the selected cell groups.

Other functions: click on the extend button, other functions will be shown as the figure below. Operate as the name of the button according to your need.

Prediction：

This function is developed for users to predict cell groups of interest in 2D or 3D view (tSNE or UMAP). Then the marker gene list and the inferred cell type will be calculated based on the CellMarker database. For cell types in the CellMarker database, we can find the top entries that are significantly enriched in the given marker list compared to all genes in human or mouse genome, under the null hypothesis that the genes of the given marker list were randomly sampled from the genome (no enrichment found). The less the p-value is, the more likely the cell type will be the predicted one.

To predict the cell type, select a cell cluster by clicking on the Predict button on the right side of the legend. The pop-up window includes the Marker Gene tab and Predicted Cell types tab. Marker Gene tab shows the Marker Gene table of selected cell cluster. The logFC represents log2(fold change). Users can filter these genes by Expression or Cell count, and then Use these genes for gene set analysis will lead you to the Gene Set Analysis Page and do an enrichment analysis with these genes. You can also switch to the Predicted Cell Types tab, which shows possible cell type prediction.

Edit: Edit the cell cluster name;
Delete：After clicking on the delete button, all the deleted cell clusters color will be turn to grey and names will be changed to "undefined" automatically. The undefined cell cluster will not be used in subsequent analysis.

# 4.2.4 Top-right Tool Bar

Use scroll of the mouse to zoom in or zoom out the dimensional reduction plot when the mouse hovers on the plot.

Hide sider: Click on the Hide sider button to hide the parameter panel to view the plot.
Pan: Click on Pan button to move the whole plot in the canvas.
Selection (select-recluster): select a bunch of cells with Box select or Lasso select tool, then a menu will pop up. Click on Cancel to go back to the plot.
Assign Selection: Assign the selected cells to a cluster of one group of customized annotation. Select the customized group from the drop-down list of Category Name and then select the cluster name.
Recluster Selection: Click on Recluster Selection to re-cluster the selected cells. This function is same as the “Recluster” function above.

Zoom: Click on the Zoom button and select a bunch of cells on the canvas to zoom in.
Zoom In: Click on Zoom In button to zoom in the whole plot.
Zoom Out: Click on Zoom Out button to zoom our the whole plot.
Reset Scale: Click on Reset Scale to go back to the default visualization of the plot.
Save as image: Export the current visualization of the canvas as a png-formatted file.
Add to report button is in the upper right corner of each downloadable image. Click the button to add the image to the report.

# 4.2.5 Expression Module

The gene expression level can color the embedding result. Use this function with marker genes would help distinguish the cell type of each cluster.

Click on the View on the Plot button in the Expression Module, the average expression of the five default genes shown below the input box will show up in the plot. The default genes can be deleted by clicking on the button on the right side of the gene or Delete all button. Input and select one or several genes then click on Add to List button, embedding plots with each gene expression level will show up in Matrix View when selected.

The gene expression level can also color 3D embedding plots. Click on 3D view, Input and select one or several genes then click on Add to List button, then the 3D embedding plot will be colored by the average expression level of your inputted gene(s).

Cell×Gene Expression Matrix Download Service is also available for single dataset. Gene expression on specific cells can be downloaded as matrices with selected cell meta information for downstream analysis. This function can be unlocked by purchasing Data Quota and is now available on both Search Gene Page and Dimensional Reduction Page of each single dataset.

Data Quota of each cell×gene expression matrix to be downloaded = Number of Selected Genes × Total Cell Count.

Selected cell meta will be downloaded together with expression matrix and will not cost any Data Quota.

Download history is shown in Download Expression Matrix Page, on which the download of expression matrix of search gene result will also be shown in the list. The detailed information of each task can be shown when hovering on the corresponding item. Matrices can be repeatedly downloaded without extra deduction of Data Quota. Used Data Quota and Total Data Quota are shown on the top-right corner of the list, and can also be found in account information by hovering on the icon in the navigation bar.

# 4.2.6 Full Screen

Full screen function is added on the upper right corner, it enlarges the visualization canvas for a better presentation.

# 4.2.7 Left-side Tool Bar

Plot Theme: Click on Plot Theme button to change the background color of the website. The background color is white by default and can be switched to black.
Save Report: Click on Save Report icon to download all the images added to report in batch.

# 4.2.8 Cell QC metrics

Cell QC metrics can be shown on the dimensional reduction plot. Click on the View on the Plot button in the Cell QC Metrics Module, and three metrics can be chosen: Number of Total Counts, Number of Genes and Mito Percent.

Choose one QC metric to color the dimensional reduction plot. The metric for each cell can be shown when the mouse hovers over the point in the plot. The cell QC metrics are only available when the raw count matrix is provided. When the raw count matrix is not provided, the dimensional reduction plot will be colored in black when Cell QC Metrics view is selected.

# 4.3 Paga

Choose the Paga Page to show the PAGA plot. PAGA plot shows the connectivity between clusters, and each node represents a cell group, differentiated by their colors, with node size indicating the number of cells, the width of edges reflecting connectivity between groups.

Filter groups: groups can be filtered via clicking on the legend button on the right.
Min connectivity threshold: connectivity threshold can be adjusted to show a different level of connectivity.
Expression：Cell cluster nodes on the PAGA plot can be colored by the gene expression via the Search button.

# 4.4 Differential Expression

Differential Expression Page facilitates to explore the gene expression difference between cell groups and subgroups. Box Plot, Violin Plot, Heatmap and Bubble Plot are available on this page.

# 4.4.1 Box Plot

Box plot reflects the differential expression of the gene or gene set across different cell groups. F-test is used for each gene or gene set to identify whether it is differentially expressed across all cell groups simultaneously, under the null hypothesis that the gene or gene set does not differentially express among all cell groups. The less the p-value is, the more likely the gene or gene set was differentially expressed. To determine whether the gene or gene set expression level of a group is different from the rest groups, OmniBrowser executes a T-test between a selected group of cells (experiment group) against the other cells (control group). The less the p-value is, the more likely the gene or gene set will be differentially expressed in this group.

Click on Differential Expression and in Box Plot tab shows the box plot.

Generate box plot: Input and select the target gene or gene set, or click on the icon under the input box to open data view, which can be more convenient to copy and paste genes in batch. Then click on Search, you may also change Group By or SubGroup By properties to adjust the box plot. Subgroup function highlights the differential expression among subgroups. Group By can select the first-layer grouping. the second-layer grouping can be selected with the SubGroup By button and the number of subgroups you need can be determined in SubGroup In item. Finally, subgroups can be defined in the drop-down list.

The average expression of all input genes will be shown as default. Click on Single gene of Plot By to switch to box plots of each gene.

Groups can also be filtered from the legend on the Groups tab.

F-test is conducted on the fly by default to evaluate statistical significance. The overall F and p values can be found on the top of a box plot. T-test result of each group compared with the rest shows up when your mouse hovers over the group bar.

For datasets with a large number of groups, the full plot is too dense to show each group clearly. Now we have added a slider at the bottom to zoom in for a selected range of the plot.

# 4.4.2 Violin Plot

Violin plot reflects the expression distribution of the gene or gene set across different cell groups. F-test is used for each gene or gene set to identify whether it differentially expresses across all cell groups simultaneously, under the null hypothesis that the gene does not differentially express among all cell groups. The less the p-value is, the more likely the gene was differentially expressed. To determine whether the gene or gene set expression level of a group is different from the rest groups, OmniBrowser executes a T-test between a selected group of cells (experiment group) against the other cells (control group). The less the p-value is, the more likely the gene or gene set will be differentially expressed in this group.

Click on Differential Expression and in Violin Plot tab shows the violin plot.

Generate violin plot: Input and select the target gene or gene set, or edit genes in data view, and then click on Search, you may also change Group By properties to adjust the violin plot. Groups can also be filtered from the legend on the Groups tab.

F-test is conducted on the fly by default to evaluate statistical significance. The overall F and p values can be found on the top of a violin plot. T-test result of each group compared with the rest shows up when your mouse hovers on the group bar.

Subgroup function highlights the differential expression among subgroups. Group By can select the first-layer grouping. the second-layer grouping can be selected with the SubGroup By button and the number of subgroups you need can be determined in SubGroup In item. Finally, subgroups can be defined in the drop-down list.

The average expression of all input genes will be shown as default. Click on Single gene of Plot By to switch to violin plots of each gene.

# 4.4.3 Heatmap

Gene-cluster heatmap displays gene expression patterns across clusters. Each unit of the matrix represents the average gene expression within the cell group.

Click on Differential Expression and in Heatmap tab shows the Heatmap.

Generate heatmap: Input and select the target gene or gene set, or edit genes in data view, and then click on Search, you may also change Show Type from all Clusters to All Cells to adjust the heatmap. Groups can be altered from the drop-down list in the Group By parameter.

Dendrogram: the dendrogram on the heatmap shows the similarity between groups and between genes. The shorter the link that connects two elements represents the more similarity.

Data used to generate the heatmap can be exported via Export button.

# 4.4.4 Bubble Plot

Box plot reflects the differential expression and the expressed fraction of every single gene across different cell groups.

Click on Differential Expression and in Bubble Plot tab shows the Bubble Plot.

Generate Bubble Plot: Input and select the target gene or gene set, or edit genes in data view, and then click on Search, you may also exchange the visualization of expression and expressed fraction to adjust the bubble plot.

Groups can be altered from the drop-down list in the Group By parameter.

# 4.5 Correlation

The Correlation Page offers scatter plot, pairwise correlations heatmap, gene-wise similarity graph and similar genes query tool to explore the relationship between genes and group of genes.

# 4.5.1 Scatter plot

The correlation scatter plot is used to visualize the association between genes or gene sets. The Pearson Correlation Coefficient and p-value can be found on the title of the plot.The p-value was calculated under the null hypothesis that the expression of such gene set signatures are drawn from independent normal distributions. The less the p-value is, the stronger the correlation is.

Click on the Correlation Page and in Scatter Plot tab shows the pairwise correlation scatter plot.

Generate scatter plot: Input and select the target gene(s) for Signature A and Signature B, choose Normal as Plot Type, then click on Search. Groups can also be filtered from the legend on the right.

Show FACS: Click on Yes to Show FACS by adding reference lines in this plot, you can drag the lines to explore accurate cell number or proportion of the divided four regions. Click on No will add a linear regression curve and equation in this plot.

The density plot displays four subplots within one graph. Besides the original scatter plot, the two histograms represent the expression distribution of signature A and signature B separately, while the contour graph shows the density distribution of the two signatures.

Switch to density scatter plot: normal scatter plot can be toggled to density plot by Density button.

# 4.5.2 Pairwise Correlations

The correlation heatmap is used to display gene similarity and cluster similarity.

Click on the Correlation Page and in Pairwise Correlation tab shows the pairwise correlation heatmap. Switch the plot type by clicking on Heatmap or Graph. You can visualize the similarity by Pearson Correlation Coefficient or Jaccard Index.

Gene-wise Similarity Graph: Click on Graph to show the similarity between genes or clusters. Each node represents for a gene or a cell group, and the width of edges reflecting the similarity between groups.

Change and filter groups: groups can be altered and filtered via the drop-down list on the right.

Min connectivity: connectivity threshold can be adjusted to show a different level of connectivity.

Dendrogram: the dendrogram on the heatmap shows the similarity between groups or between genes. The shorter the link that connects two elements represents the more similarity.

Pairwise Correlation Heatmap: You can input the gene set and select the group in the drop-down list of Group By and then clusters you want to include, then toggle the comparison unit between cluster and gene from Compared by parameter. If Compared by Gene is chosen, then you may further define the Calculated By parameter.

Data used to generate the heatmap can be exported via Export button.

# 4.5.3 Similar genes

Similar Genes tool queries all the similar genes of the selected gene. Pearson Correlation Coefficient is calculated between the user-input gene and each of the other genes, in order to find the most correlated genes. The p-value was calculated under the null hypothesis that the expression of such gene set is drawn from independent normal distributions. The less the p-value is, the stronger the correlation is.

Click on the Correlation Page and in the Similar Genes tab, you can input and select your target gene.

You may also choose different groups and include clusters you want, then click on Search to view the result table. The result can be sorted by Pearson Correlation Coefficient or p-value. Ligand and receptor info are also listed if any. Data in the table can be exported via Export button.

# 4.6 Marker Genes

The Marker Genes Page shows marker gene information of a cluster, also allows users to verify the identity of certain clusters and help identify any unknown clusters. To determine each gene is a marker gene or not, a T-test or Wilcoxon-test is used to identify whether a gene differentially expresses between a selected group of cells (experiment group) against the other cells (control group). The less the p-value is, the more likely the gene will be a marker gene.

Click on the Marker Gene Page shows the marker gene list query function.

# 4.6.1 Marker Gene List

Select a meta group in Meta Name and cluster(s) of interest. Choose One vs Rest or Group vs Group as the way of marker gene calculation. To filter positive or negative marker genes, please choose Statistic to be LogFC first and then select by Marker Type. Click on Search shows the marker gene table of the selected cluster(s). LogFC represents the log of fold change.

Hover over the gene ID to show the external links of Search Gene Product and Show Related Genes. Clicking on Search Gene Product will jump to the website of AmiGO 2 showing the gene and gene products. Clicking on Show Related Genes will show the ideograms with interacting genes and paralogous genes. The External Data tab shows the distribution of pathogenic-related genes recommended by ACMG, which can provide more references for users to do gene research.

The table can be exported to excel files via the Export button. You can click on Use these genes for gene set analysis button to carry out gene set analysis.

# 4.6.2 Volcano Plot

Click on the Marker Gene page and in Volcano Plot tab shows the volcano plot. Choose test method, meta group by Meta Name and cluster, calculation type to visualize. Positive marker genes are colored red and negative marker genes are colored green. Turn off the coloring by switching Coloring. Input genes in Gene Set box to display the gene name on volcano plot, the tag of each gene can be dragged on the plot to make a better visualization and export as a png-formatted image. Set the horizontal cutoff line by -log10 Qvalue and the vertical cutoff line by log2FC.

# 4.7 Supervised Annotation

Supervised annotations utilize the SciBet algorithm and the public dataset to annotate your dataset. For each dataset, a reference kernel was trained by the SciBet method, so that this dataset could be used as a reference to perform cell type annotation on uploaded scRNA-seq data. Before you run this function, please choose a dataset most similar to your data in cell type. You may view the dataset profile on the Load Dataset Page of each dataset.

# 4.7.1 Cell-type Composition

Cell-type Composition tab shows all the predicted cell types distribution of your uploaded cells.

On the Supervised Annotation page, click on Choose File to upload your tsv file then you can see the predicted annotation results of your data. A sample dataset is available via the Download test dataset button. You may refer to its format and attributes.

# 4.7.2 Sankey Plot

Click on the Sankey Plot tab shows the Sankey plot of the uploaded data. Sankey Plot tab shows the one-to-one correspondence of your annotation and the predicted annotation. On the left is the annotation from the cellType attribute of your uploaded file, the right side gives the SciBet algorithm predicted annotation.

# 4.7.3 Classification Result

Click on the Classification Result tab shows the predicted cell type. Classification Result tab shows the predicted cell type of each cell uploaded. The result table can be downloaded via the Export button.

# 4.7.4 Classification Probability

Click on the Classification Probability tab shows the predicted probability of each cell type from the reference dataset at the single-cell level. Classification Probability tab shows the predicted probability of each cell type. The result table can be downloaded via the Export button.

# 4.8 Metadata

Cell metadata visualizes the relationship between two groups of meta information of this dataset.

# 4.8.1 Bar Plot

Click on the Metadata Page and in Bar Plot tab shows the bar plot of two groups of metadata. Choose one group of metadata in Group By drop-down list and another in Color By to explore two different properties. The two groups can be swapped by clicking on the button of Swap Axis. The visualizing two properties can be filtered on the right. The plot can be shown according to cell number or percentage of features. The legend of the data can be hidden by clicking on Hide. Listed information displays when the mouse hovers over a bar.

# 4.8.2 Heatmap

Click on Metadata page and in Heatmap tab shows the metadata heatmap. Choose one group of metadata in Group By drop-down list and another in Color By to explore two different properties. The two groups can be swapped by clicking on the button of Swap Axis. The percentage in row displays when the mouse over a bar. The dendrogram on the heatmap shows the similarity between groups and between genes. The shorter the link that connects two elements represents the more similarity.

# 4.8.3 Sankey Plot

Click on the Sankey Plot tab shows the Sankey plot of the metadata. Sankey Plot tab shows the one-to-one correspondence of one meta group and another meta group. The two groups can be swapped by clicking on the button of Swap Axis.

# 4.9 Gene set Analysis

Click on Gene Set Analysis page to do the gene set analysis. Gene Set Analysis Page provides various ways to analyze the pathways in which the input gene set of interest participate. There are three tabs in Gene Set Analysis Page: Gene Ontology tab, WikiPathways tab and Reactome tab.

# 4.9.1 Gene Ontology Analysis

Gene Ontology tab is used to query GO enrichment results. For GO terms in the gene ontology database, we can find the top entries that are significantly enriched in the given gene list compared to all genes in the human or mouse genome, under the null hypothesis that the given genes were randomly sampled from the genome (no enrichment found). The less the p-value is, the more likely the entry will be the enriched one.

Gene set can be inputted and selected in the input box or click on the icon under the input box to open data view, which can be more convenient to copy and paste genes in batch. Data view supports text files with one gene per line or genes separated by various delimiters. View detail list button on Load Dataset page and Use these genes for gene set analysis button on the Marker Genes page may also import the result gene list to this function.

The results can be filtered or sorted through buttons below the gene set input box. Please note Species must be chosen correctly according to the gene set queried to get a reasonable conclusion. Namespace includes three types: molecular function(MF), cellular component(CC) and biological process(BP).

Click on a GO item will lead you to the related page on QuickGO.

Click on Bar Plot in Plot module to view the bar plot of GO Functional Analysis. This bar plot can be exported and saved as png-formatted image.

# 4.9.2 WikiPathways Analysis

WikiPathways tab is used to query pathway information, based on the pathway database WikiPathways. We can locate the particular pathways by inputting a gene list of interest in this tab.Each pathway will be provided with a p-value. The less the p-value is, the more likely the pathway will be the enriched one. The result can be further searched by pathway ID or pathway name via the search box on the top left of the table.

The results can be filtered or sorted through buttons below the gene set input box. Please note Species must be chosen correctly according to the gene set queried to get a reasonable conclusion.

Click on a Pathway ID will lead you to the related page on WikiPathways.

Click on Bar Plot in Plot module to view the bar plot of WikiPathways Functional Analysis. This bar plot can be exported and saved as png-formatted image.

# 4.9.3 Reactome Analysis

Reactome Analysis tab is used to query signaling and pathway enrichment results, based on the database Reactome. Entities participating in reactions form a network of biological interactions and are grouped into pathways, and we can locate the particular pathways or reactions by inputting a gene list of interest in this tab. Each pathway will be provided with a p-value. The less the p-value is, the more likely the pathway will be the enriched one. The result can be further searched by pathway ID or pathway name via the search box on the top left of the table.

The results can be filtered or sorted through buttons below the gene set input box. Please note Species must be chosen correctly according to the gene set queried to get a reasonable conclusion.

Click on a Pathway ID will lead you to the related page on Reactome.

Click on Bar Plot in Plot module to view the bar plot of Reactome Functional Analysis. This bar plot can be exported and saved as png-formatted image.

# 4.10 Immunoreceptor

Some datasets include immunoreceptor information. To view these datasets, select Immnunoreceptor in Cell Meta tab of Search Page and apply the filter, datasets with immunoreceptor information will show up. Immunoreceptor information can be viewed via TCR tab in Dimensional Reduction Page and via TCR page in a single dataset.

# 4.10.1 View Immunoreceptor Info On Embedding

Click on View on the Plot of TCR tab on Dimensional Reduction Page to visualize TCR on embedding plot. Clonotype size in all clusters will be displayed as default. To view the clonotype size in specific clusters, select the cell cluster(s) to be displayed in the Category view, and then click TCR view and Calculate clonotype size in Selected clusters to calculate and display the clonotype size in the selected clusters.

Input the Clonotype ID(s) in the input box and choose clonotype ID(s) in the drop-down list, and then click on Add to List button to view the corresponding immunoreceptor distribution.

Click on Plot By Clonal Expansion, input high-cutoff of the clonotype size and click on Submit to show the enrichment of TCR on embedding. Click on Reset to clear number input.

# 4.10.2 Clonal Type

Clonal Type ID, size, expansion and detailed sequencing data will be displayed as a table in Clonal Type tab. Choose one filter and input keywords to filter the table by clicking on Apply Filter in the parameter panel. Cell annotation label can be switched by the drop-down list of Cell Type, and the Clusters shown below will be changed correspondingly.

To view the clonotype size in specific clusters, select the cell cluster(s) in Clusters, and then click on Calculate clonotype size in Selected clusters to calculate the clonotype size in the selected clusters. The table can be exported as excel formatted file by clicking on Export button on the top right corner of the table area.

# 4.10.3 Clonotype Abundance

Click on the TCR page and in Clonotype Abundance tab shows the bar plot of TCR info and metadata. Choose one group of metadata in Group By drop-down list and one group of TCR info in Color By to explore two different properties. The axes can be swapped by clicking on Swap Axis. The plot can be shown according to cell Number or Percentage of features. Listed information displays when the mouse hovers over a bar.

# 4.10.4 Gene Usage

Click on Gene Usage tab to view the V(D)J recombination. The plot gives the pairing results of VDJ pairs of cells with immunoreceptor. The more the pairing between two genes, the greater the bandwidth is. Change the value of Expansion to adjust the plot. Some cells may have a second chain of TCR, and this part of information is shown in Chain 2. Click on Chain 2 to switch the type. Cell annotation label can be switched by the drop-down list of Cell Type, and the Clusters shown below will be changed correspondingly. Choose clusters of interest to view the gene usage in the plot. The Gene Usage plot can be downloaded or added to report by the toolbar on the top right corner.

# 4.11 Deconvolution

Deconvolution module has been enabled to some datasets. Based on the inferred cell proportions in TCGA/GTEx bulk-RNA sample with a scRNA dataset as reference, the proportion comparison and cell type-level differential expression can be performed. The general principle of deconvolution is shown below in the figure derived from GEPIA2021 (opens new window). The deconvolution function is developed using MuSiC (opens new window).

The reference for MusiC(Multi-subject Single Cell deconvolution):

Bulk tissue cell type deconvolution with multi-subject single-cell expression reference X. Wang, J. Park, K. Susztak, N.R. Zhang, M. Li Nature Communications. 2019 Jan 22 https://doi.org/10.1038/s41467-018-08023-x (opens new window)

Only some human datasets enable deconvolution function now. Select Deconvolution in Study Meta tab of Search Page and apply the filter, datasets with deconvolution function will show up.

Proportion Analysis: Visualize the proportion of each cell type selected with the interactive boxplot. Users can perform the quantitative comparison (ANOVA) of the proportion among cell types or TCGA/GTEx sub-datasets.

Sub-expression Analysis: Visualize the gene expression in each cell type selected with the interactive boxplot, and perform the cell type-level differential expression analysis. Similar with the Proportion Analysis, the differential expression analysis with ANOVA is also available.

# 4.12 Spatial Transcriptome

Spatial Transcriptome data assign cell types to their locations in the histological sections. Only some of 10× official datasets contain spatial transcriptome data now. Select spatialTranscriptome in Study Meta tab of Search Page and apply the filter, datasets with spatial transcriptome data will show up.

View spatial transcriptome data: Click on Spatial on the right side of Dimensional Reduction Page. Spatial information can also be related to different meta information on spatial plot. Click on some specific cell type to show its spatial distribution only. Specific gene expression level can also be displayed according to spatial distribution. Input and select genes in the expression box and click on View on the Plot to show.

# 5. Cross-Dataset Analysis

Dataset Cross Analysis function is used to compare the different datasets with visual interface. It supports to compare up to 4 datasets with the same gene expression inputted. Gene expression, cell groups, cell clusters, metadata and Marker genes of each dataset can be viewed together to show the difference.

In Search Page, the datasets can be added to cart or removed. Clear Selection button is used to clear all the datasets in cart. And Cross Analysis is used to compare the selected datasets.

In the Cross Analysis Page, the selected datasets are in the low part of screen, after selecting the datasets and clicking the Apply Selections button, the compared result can be viewed by Dimensional Reduction, Differential Expression, Metadata and Marker Gene function;

# 5.1 Dimensional Reduction

# 5.1.1 Projection

You can visualize the dimensional reduction plots only in 2D view in Cross Analysis. Methods can be toggled between tSNE and UMAP.

Category View: Plot can be colored by different groups of annotation labels, e.g. Cluster Name, Cell ontology name.

# 5.1.2 Category

All group labels in the applied datasets are in the drop-down list here. For different datasets, if the cell cluster names are identical, the cell group will added together. If there is unique cell group in one dataset, the other datasets don’t show the dimensional reduction plot. And all the cell clusters can be sorted by the cell count or the group name.

Above the cell cluster name, there are 4 buttons from left to right:

Notification: when the mouse hovers over the button, it will show a notification.
Search: input the cell cluster name to search cell cluster;
Save as image: save the category part as png-formatted image;
Add to report: add the category image to the report;

There is the legend of the plot.The representing pattern of each dataset displays above the legend of the data. Click on the check box on the left of the cell cluster name, the selected cell cluster will be hidden; When your mouse hovers over one cell cluster, the component of this cluster will be showed, it shows the cell number and fraction of each dataset. Click on the cell cluster to highlight the cluster in the plot. Click again to remove the highlight. Also, the color of each group is auto-generated to ensure that the color contrast of adjacent points in the chart is big enough for a good presentation.

# 5.1.3 Top-right Tool Bar

Hide dataset list: Click on the Hide dataset list button to hide the dataset list module for a better view of the plot.
Hide parameter panel: Click on the Hide parameter panel button to hide the parameter modules for a better view of the plot.
Reset Scale: Click on Reset Scale to go back to the default visualization of the plot.
Save as image: Export the current visualization of the canvas as a png-formatted file.
Add to report button is in the upper right corner of each downloadable image. Click the button to add the image to the report.

# 5.1.4 Expression

The gene expression level can color the embedding result. Use this function with marker genes would help distinguish the cell type of each cluster.

Click on the View on the Plot button in the Expression Module, the average expression of the five default genes shown below the input box will show up in the plot. The default genes can be deleted by clicking on the button on the right side of the gene or Delete all button. Input at least one gene and click Add to List, each dimension reduction plot of datasets will show the average gene expression, if the plot is all dark-grey, it means that no gene expresses on this plot. When clicking the Batch Correction option is Yes, batch effect among these datasets will be removed. The Batch Correction option only supports to remove the batch effect of genes which all the datasets contain in the analysis.

# 5.1.5 Left-side Tool Bar

Plot Theme: Click on Plot Theme button to change the background color of the website. The background color is white by default and can be switched to black.
Save Report: Click on Save Report button to download all the images added to report in batch.

# 5.2 Differential Expression

Differential Expression Page of Cross Analysis facilitates to explore the gene expression difference between datasets. Violin Plot and Bubble Plot visualizations are available on this page.

# 5.2.1 Violin Plot

Violin plot reflects the expression distribution of the gene or gene set across different cell groups of different datasets. F-test is used for each gene or gene set to identify whether it differentially expresses across all cell groups simultaneously, under the null hypothesis that the gene does not differentially express among all cell groups. The less the p-value is, the more likely the gene was differentially expressed. To determine whether the gene or gene set expression level of a group is different from the rest groups, OmniBrowser executes a T-test between a selected group of cells (experiment group) against the other cells (control group). F-test is conducted on the fly by default to evaluate statistical significance. The overall F and p values can be found on the top of a violin plot.

Generate Violin Plot: Input genes and click Add to List and explore the differential expression among datasets. Use batch correction function if needed. The violin plot can be visualized according to cell category or dataset, click on X** Axis** to switch the view. Groups can also be filtered from the legend on the Groups tab. Groups can be altered from the drop-down list in the Group By parameter.

# 5.2.2 Bubble Plot

Bubble Plot reflects the differential expression and the expressed fraction of every single gene across different cell groups from different datasets.

Click on Bubble Plot button in Plot Type to show Bubble Plot.

Generate Bubble Plot: Input and select the target gene or gene set then click on Search, you may also exchange the visualization of expression and expressed fraction to adjust the bubble plot. The gene expression and gene expressed fraction can be set to filter the cell group.Groups can be altered from the drop-down list in the Group By parameter. For comparing the gene expression in each dataset, there is also batch correction option to remove the batch effect(for details of batch correction, please refer above).

# 5.3 Metadata

Cell metadata outlines the overall information across datasets.

# 5.3.1 Bar Plot

Click on the Metadata page and choose Bar as Plot Type shows the bar plot of two groups of metadata. Choose one group of metadata in Group By drop-down list and another in Color By to explore two different properties. The visualizing two properties can be filtered on the right. The plot can be shown according to cell number or percentage of features. The legend of the data can be hidden by clicking on Hide. Listed information displays when the mouse hovers over a bar.

# 5.3.2 Heatmap

Click on the Metadata page and choose Heatmap as Plot Type shows the bar plot of two groups of metadata. Choose one group of metadata in Group By drop-down list and another in Color By, then choose one specific cell cluster to explore two different properties of the specific cell cluster. The dendrogram on the heatmap shows the similarity between groups and between genes. The shorter the link that connects two elements represents the more similarity.

# 5.4 Marker Gene

The Marker Gene Page shows marker genes of different clusters from different datasets. To determine each gene is a marker gene or not, a T-test or Wilcoxon-test is used to identify whether a gene differentially expresses between a selected group of cells (experiment group) against the other cells (control group). The less the p-value is, the more likely the gene will be a marker gene.

Search Marker Gene: Select a meta group in Meta Name and cluster(s) of interest. To filter positive or negative marker genes, please choose Statistic to be LogFC first and then select by Marker Type. Every click filters the result instantly. Marker genes of specific dataset or cluster can be filtered by clicking on the title of the marker gene table. LogFC represents the log of fold change.

Visualize via Bubble Plot: select the target gene or gene set then click on Plot. There are up to 10 marker gene can be selected on current page, and double spread selection is prohibited. Click Plot button and select the batch correction option to generate the bubble plot. The gene expression and gene expressed fraction can be set to filter the cell group. When parameters on the right are modified, the new plot needs to be re-plotted.

# 6. Upload Dataset

Upload Dataset function is authorized to certain users. Users with permissions can find the two services on the sidebar. If you are interested in these functions, please contact us.

Upload Dataset page gives the user access to upload their datasets to OmniBrowser. You can easily compare your own data with published papers after uploading while still keep your data private.

Click on Upload Dataset button to go to Upload List, which shows your previous uploaded dataset list.

Click on the title of an uploaded file leads you directly to the dataset. Upload status will show the current process status after you submit a dataset.

The present permissions of each dataset displayed below the Permissions feature.

You can use the Delete button to remove an uploaded dataset. Only the uploader has access to delete the dataset.

You can easily manage data view and download permissions after you upload a dataset here.

If you want to change the release permission, you can click on the Permissions button in the Operations column. A window will show up for you modify the authority. For batch update permissions, please click on the dataset checkboxes then click on the Batch Update button.

Here you can set View permission and Download permission separately. View permission gives users access to explore visualizations and analyze results of the uploaded dataset. Download permission offers users access to download the uploaded dataset in h5ad format.

Four levels of authority are provided: private means only the uploader have access to this dataset, company means users from the same company have access to this dataset, department means users from the same department have access to this dataset, public means all OmniBrowser users have access to this dataset.

To upload a dataset, you need to click on the Upload new button on the Upload Dataset Page, then fill in the Upload New Dataset sheet and upload your file.

In the Upload New Dataset sheet, Title, Tissue, Species, Research topic, and Library method must be filled. After clicking on Submit, you will be redirected to the Upload List Page. Your dataset will be ready to explore when the status change to Success.

Currently, HDF5 file and TPM file formats are supported, please read the hint and help information on this page to ensure compatibility. If you need to upload a TPM file, please choose the Upload file type to be Other. If you upload the Marker gene file, please choose a test method before submitting it.

TPM format examples are provided via the Download example button.

You can use the dataset type in the upper search box to view different datasets. The public dataset includes datasets curated by our data curation team. The uploaded dataset includes datasets uploaded by users.

# 7. Download Dataset

Download Dataset function is authorized to certain users. Users with permissions can find the two services on the sidebar. If you are interested in these functions, please contact us.

The Download Dataset shows the datasets user have access to download.

To download a dataset, click on the Download button on the right of each dataset.

Each dataset has five-time download permission for each user. Downloads Remained shows how many times are left for each dataset.

In the List view of Search Study, a dataset with download permission will also show a download button on the right. Please contact us to apply for download permission if you are interested.

# 8. Reference List

HPA:The Human Protein Atlas http://www.proteinatlas.org/ (opens new window)
HCL:Human Cell Landscape http://bis.zju.edu.cn/HCL/ (opens new window)
Scibet (Supervised Annotation) https://www.nature.com/articles/s41467-020-15523-2 (opens new window)
Tabula Muris https://tabula-muris.ds.czbiohub.org/ (opens new window)，DOI: 10.1038/s41586-018-0590-4 (opens new window)
Tabula Sapiens DOI: 10.1126/science.abl4896 (opens new window)
Pegasus (underlying algorithm for markerGeneAnno) https://github.com/lilab-bcb/pegasus (opens new window), reference publication https://pubmed.ncbi.nlm.nih.gov/32393754/ (opens new window)
UBERON:Uber-anatomy ontology https://www.ebi.ac.uk/ols/ontologies/uberon (opens new window)
Disease Ontology https://disease-ontology.org/ (opens new window)
Cell Ontology https://www.ebi.ac.uk/ols/ontologies/cl (opens new window)
MuSiC:Multi-subject Single Cell deconvolution https://xuranw.github.io/MuSiC/index.html (opens new window)
GO Enrichment Analysis http://geneontology.org/ (opens new window)
Gene-NCBI (marker gene link) https://www.ncbi.nlm.nih.gov/gene/ (opens new window)
Related genes:Ideogram of interacting genes and paralogous genes https://eweitz.github.io/ideogram/annotations-external-data (opens new window)
Gene product: AmiGO 2 http://amigo.geneontology.org/amigo/landing (opens new window)
External data:ACMG recommendations for reporting of incidental findings in clinical exome and genome sequencing https://www.sciencedirect.com/science/article/pii/S1098360021027623?via%3Dihub (opens new window)
Wikipathway:Martens M, Ammar A, Riutta A, Waagmeester A, Slenter DN, Hanspers K, Miller RA, Digles D, Lopes EN, Ehrhart F, Dupuis LJ, Winckers LA, Coort SL, Willighagen EL, Evelo CT, Pico AR, Kutmon M. WikiPathways: connecting communities Nucleic Acids Research, (2021), Volume 49, Issue D1, 8 January 2021, Pages D613–D621, https://doi.org/10.1093/nar/gkaa1024 (opens new window) PMID:33211851 (opens new window)
Reactome:Marc Gillespie, Bijay Jassal, Ralf Stephan, Marija Milacic, Karen Rothfels, Andrea Senff-Ribeiro, Johannes Griss, Cristoffer Sevilla, Lisa Matthews, Chuqiao Gong, Chuan Deng, Thawfeek Varusai, Eliot Ragueneau, Yusra Haider, Bruce May, Veronica Shamovsky, Joel Weiser, Timothy Brunson, Nasim Sanati, Liam Beckman, Xiang Shao, Antonio Fabregat, Konstantinos Sidiropoulos, Julieth Murillo, Guilherme Viteri, Justin Cook, Solomon Shorser, Gary Bader, Emek Demir, Chris Sander, Robin Haw, Guanming Wu, Lincoln Stein, Henning Hermjakob, Peter D’Eustachio, The reactome pathway knowledgebase 2022, Nucleic Acids Research, 2021;, gkab1028, https://doi.org/10.1093/nar/gkab1028 (opens new window)