This manual will guide you through the process of using TIP as an in silico web server to track, analyze and visualize the status of anti-cancer immunity and the proportions of tumor-infiltrating immune cells across seven-step Cancer-Immunity Cycle using RNA-seq or Microarray expression data.
We display a vivid portrayal of the biological process of 7 steps on the home page as shown below.
1. Click on "Start to analysis..." in the middle of The Cancer-Immunity Cycle, there will pop-up one page to import files and fill in parameters.
2. Click on "Analysis" at the top right corner of the home page. Produce the same effect with above.
3. Click on "PancancerAnalysis" at the top right corner of the home page to start analysis the 11373 pan-cancer samples for 33 types of cancer from TCGA. there will pop-up one page to select the cancer type you are interested in.
4. When the job runs, a specific job ID will generate immediately. Fill in your job ID in the search box to check history job results or click on "Example" to see the example job ID. It is recommended that users download their results as soon as possible after the job is done, as the results of a job will only be stored on our server for one months.
5. The purpose and main function of TIP.
6. The statistical information of TIP.
"Analysis" performs a systematic analysis of tumor immunophenotype, including the status score of anti-cancer immunity and proportions of tumor-infiltrating immune cells by user uploading gene expression datasets.
The first step is to prepare your tumor expression data files if you want to analysis your own expression data. The input files should be in text tab-delimited format with no double-quote marks. In general, the expected file format is a table where the first row consists of column headers containing the sample labels and the first column consists of row headers containing the gene identifier (Gene Symbol, Gene ID, or Ensembl ID without version number), with the data points occupying the remainder of the table. There should be no missing values within the table. We suggest you upload expression profile with no more than 50 samples in case memory overflow at permuting. The results of seperated samples can be merge directly. You can upload two types of data:
Microarray data: Different platforms are available, such as, Illumina or Agilent, etc.
Below is a snapshot of Microarray example data:
RNA-seq data: TPM and count data type is allowed. In general, you’d better use TPM data type, as the TPM value is an ideal data type for RNA-seq measurements. even so, we can also convert count to TPM value during the file validation step at the cost of a little time and some genes.
Once you have prepared your data files, you may upload them to the TIP website, which will appear like shown below.
1. Upload files: Click "Browse" to upload your own files.
2. Cancer type: Select a cancer data to be the reference background from the drop-down box, which contains cancer and normal samples to compare with your samples. There are 33 alternative cancer types.
3. Type of data: TIP support two kinds of data, RNA-seq and Microarray. Note: RNA-seq TPM data should not be log-transformed.
4. Format of file: If you choose RNA-seq data types, you will have two kinds of data format to select, TPM or Count. If you choose Microarray data type, ignore it.
5. Email (optional): Input your own email address, we will send your results to your mail box.
6. Click the "Submit" button, TIP will generate results in minutes (1.2 minutes per sample on average).
Note: Here, RNA-seq of TPM data was used extensively in benchmark analysis in the companion manuscript, will be used to demonstrate the workflow for TIP. If you wish to go straight to running the example datasets and go through the manual later, you may run it as an example dataset.
Instruction: At the top of the browse box, click on "example data" (blue), and you will see the data name in the box that shown below. the others parameter is filled in by default, then submit it directly. There is also an example download link nearby.
After submitting your data, the TIP is started analysis automatically. On the top of the pop-up page, there is a frame of job information.
Note for some signs: Click the sign at the top left-hand corner of some frames to check the detail information of this section. As follows:
Click the sign in the top right-hand corner of the frame to download result data of this section. As follows:
Click the sign in the top right-hand corner of the picture to download the picture for different formats. As follows:
TIP’s user interface after data upload looks as the following three visual outputs, produced for whole samples.
An example of generated heatmap, using the available sample data, is shown in figure 1. This plot provides an overview of immune activity score for TCGA tumor data observed in the seven steps of the Cancer-immunity Circle.
Figure 1: The heatmap shows the 23 activity scores (row names) of anti-cancer immunity across seven-step Cancer-Immunity Cycle for all samples (column name). The color of the bars indicates the activity score of each sample, which means that deeper of red colors correspond to higher score as well as higher the activity of immune status. The number of genes which are annotated in the reference genes is shown in the row names.
The second visual output is stacked plot, using the available sample data, is shown in figure 2. We infer the relative proportions of 14 or 22 immune cell types for RNA-seq or Microarray data, respectively.
This stacked plot provides an overview of proportions of immune cell infiltration for all samples.
Figure 2: The interactive stacked plot shows the proportions of tumor-infiltrating immune cells of interest for all samples. Each color represents one type of immune cell. The x-axis spans the whole samples. The y-axis is the proportion of immune cell, the sum of these cells proportions is 1. Rolling the mouse over a bar gives the proportions of immune cells for this sample. Click the immune cells of interest in the legend of the plot (left), the stacked plot will change combination in real time. There are 14 types of immune cells for the example is RNA-seq data.
Note: When your data type is Microarray, there should be showing 22 types of immune cells in the stack plot.
The third visual output is combination plot, heatmap (left), and PCA scatter plot (right). Using the available sample data, is shown in figure 3.
Figure 3: Group figure of overview the signature gene expression of samples.
The heatmap (left) shows the expression levels of 178 step-specific signature genes of anti-cancer immunity across seven-step Cancer-Immunity Cycle for all samples. Each row represents a single gene, each column represents one sample. The color from red (positive) to blue (negative) in the heatmap represents expression values size down. And the seven colors’ bar at the left represent the immune steps.
Interactive principle component analysis (PCA) of signature genes expression for all samples (right). The two axes represent the two "PC" of all samples (i.e. two principal components). Data points tend to aggregate together when they are correlated. Rolling the mouse over a spot gives the value of two PCs.
In the right side of this page, there is a frame shows the global immune activity score of each sample.
1. The search box above the table is to check some samples you are interested in.
2. Under the search box is a table of samples (link) and its overall activity scores, the deeper the color, the higher the score.
3. When you want to know about the specific information of one sample, click on the sample name in blue, there will pop-up a new page about this sample, details content as follows.
4. Click the triangle sign beside "Overall activity" to sort the score.
The following three visual outputs are in the pop-up page when you choose one sample you interest:
A global group figure of overview the immunophenotype status for one sample:
Figure 4. Interactive line graph for the activity scores of anti-cancer immunity in seven steps across Cancer-Immunity Cycle for a single sample (upper left). Colors of the dot corresponding to seven steps. Rolling the mouse over a spot gives the score.
Interactive boxplot for gene expression of step-specific signature genes in seven steps across Cancer-Immunity Cycle for a single sample (bottom left). Colors of the box correspond to seven steps. Rolling the mouse over a spot gives the gene expression value.
The Interactive pie plot (right) shows the relative proportion of tumor-infiltrating immune cells for a single sample. It is also possible to zoom on and to check each immune cell.
A group figure of seven immunophenotype status score plot for one sample:
Figure 5: The series of plots show activity scores of anti-cancer immunity in seven steps for single sample comparing with reference tumor immunophenotype profiling of the corresponding cancer type from TCGA pan-cancer data.
The circular plots in step 1-3 and 5-7: There are two tracks act as the background of tumor and normal samples’ immune activity scores; One needle (red line) across two tracks points to the score of the selected sample. The green points and gray line in the red track (the outer race) represent tumor sample and mean value of scores, respectively. By contrast, the orange points and gray line in the green track (the inner race) represent normal sample and mean value of scores. In the center of the whole circle, the sample’s specific score is available. The numbers in the outermost of the circle are the scale of samples’ score. The others circular plots are similar as this.
A blown-up figure show as follows:
The radar plot in step4: It shows activity score of each immune cell in the step of trafficking of immune cells to tumors for a single sample(blue). The medians of immune cell trafficking to tumor scores from TCGA normal (green) and tumor (orange) samples are shown as references.
The radar plot in step5: It shows the relative proportions of tumor-infiltrating immune cells for a single sample (blue). The medians of tumor-infiltrating immune cell proportions from TCGA normal (green) and tumor (orange) samples are shown as references.
Rolling the mouse over a pot gives the score of sample, mean value corresponding to background cancer and normal samples can be shown by rolling the mouse over the corresponding line in the circuit. It is also possible to zoom on each figure for every step.
A blown-up figure shows as follow:
PancancerAnalysis will evaluate the status score of anti-cancer immunity and proportion of tumor-infiltrating immune cells calculated based on TCGA pan-cancer data across 33 human cancers containing 11373 samples.
You can choose one type of tumor type you interest to submit and start to analysis.
Click the "Submit" button, TIP will generate results in minutes.
Once you submit your selected cancer type, the first pop-up page is a systematic view of cancer immunophenotype status for the whole samples.
The following three visual outputs produced for whole samples during the "PancancerAnalysis" are similar with "Analysis".
While there is a difference in the PCA analysis for expression of signature genes (right). We choose BRCA as an example: Colors of dots corresponding to the type of samples, Tumor (red) and Normal (green).
If you have any questions or comments regarding to TIP, you may contact us by sending an email to Xinxin Zhang (firstname.lastname@example.org).