NEWS

15-06-2021

• Reboot's paper is now published in NAR Cancer!

15-04-2021

• Reboot's website is online! Datasets available are: BRCA, ESCA, GBM and PAAD from The Cancer Genome Atlas (TCGA).

FAQ & Help

What is "Reboot"?

Reboot is an integrative approach to find and validate not only genes, but also splicing isoforms associated with patient prognosis of several tumor types from high dimensional expression datasets. Reboot's algorithm uses a multivariate strategy with penalized Cox regression (LASSO method) combined with a bootstrap approach. In addition, Reboot provides robust statistical tests and many informative plots to support its findings.

How can I run Reboot?

You only need to choose a tumor type and the kind of analysis (gene or transcript-based) you want and hen hit 'submit'! After a few seconds, all the results (plots and files) will be ready for visualization and download using a test dataset downloaded from TCGA (add link). You can also change the coefficients filter by scrolling horizontally the bar and/or use the ROC curve to choose cutoffs for survival analysis (instead of the median). Finally, a 'validation' tab is also available in case you want to perform an independent analysis using your own datasets. Here, you are able to modify several parameteres in order to better suit the analysis for your needs.

What are the inputs?

For testing purposes, no inputs are needed because we pre-processed the data from TCGA database. If you want to validate your results and perform an independent analysis, then you should provide an expression file (.txt) like the one generated by the 'module I' of Reboot.

Which outputs can I get?

By default, 4 plots are produced:

1. Histogram with coefficients distribution. An dotted vertical line indicates the values taken according to the 'coefficients' filter;
2. Lollipop plot showing up to 20 genes/transcripts' coefficients;
3. Kaplan-Meier survival plot for the score generated by the 'module I' of Reboot with its respective p-value. A risk table is also available;
4. Forest plot showing the significance of the score adjusted for other relevant clinical parameters (according to tumor type).

If the 'ROC' option is marked, a 5th plot is drawn: a ROC curve. In addition, a plot with the proportional hazards assumptions is made (download only). Lastly, 2 '.txt' files are generated containing a summary of the results from the univariate (logrank) and multivariate (multicox) analyses.

Can I download the outputs?

Yes! All outputs generated are available for download either as PDFs (plots) or TXT (files).

Where can I find input datasets?

You can find other input datasets at TCGA.

Can I use the website to validate high-dimensional datasets?

Yes, although the performance should be slow. We highly recommend the CLI, so that you can take advantage of all Reboot's functionalities. Please refer to our detailed Documentation to Reboot your experience!

How can I cite/acknowledge Reboot?

If you use either the GUI or the CLI of Reboot, please cite us: Reboot: a straightforward approach to identify genes and splicing isoforms associated with cancer patient prognosis. Felipe R. C. dos Santos*, Gabriela D. A. Guardia*, Filipe F. dos Santos*, Daniel T. Ohara, and Pedro A. F. Galante. DOI: 10.1093/narcan/zcab024. PMID: 34316711.

What is the contact information?

You can contact Dr. Pedro A.F. Galante by email. For more information, you are welcome to visit us at our Lab's website!