Gene panels in the cBioPortal patient view

Gene panels have been part of cBioPortal for some time but were not incorporated in the patient view. The Hyve has now implemented support for gene panels in the patient view, to allow users to distinguish between alterations that were not found because they are likely absent, or because they were not profiled.


cBioPortal is a web based platform to integrate cancer genomics data from various data types (mutations, copy-number alterations, expression, clinical, etc.) to help oncologists in choosing the best treatment for their patients and researchers in uncovering the mechanisms involved in tumor development.


The Hyve manages the largest number of active cBioPortal installations in the world, for a wide variety of clients, including pharma companies, hospitals, research institutes, data providers and research collaborations. Our contributions to the open-source code base can be found in our articles.

Read more

Gene panels

To cost-effectively create a genetic profile of a tumor, gene panels are being used to focus only on a subset of genes that are often involved in tumor formation. For the data in the MSK impact study, for instance, MSK-impact panels have been used. As we learn more about oncogenes and costs of sequencing decrease, panels can be extended with additional genes, giving rise to different versions of a panel, for instance MSK-IMPACT 341 and MSK-IMPACT 400, with 341 and 400 genes respectively.

In addition to sequencing of tumor tissue, we are seeing more and more sequencing of circulating tumor DNA (ctDNA) from liquid biopsies. The earlier the stage of the tumor, the lower the concentration of the ctDNA in the blood and the higher the sensitivity that is required to pick up the signal. The higher the required sensitivity the higher the costs for the sequencing. To limit the costs, gene panels can be used to limit the number of genes. Typically, these panels contain fewer genes than the panels used for sequencing tumor tissue.

In terms of the interpretation of the data, the use of gene panels raises the complication that the absence of mutations in a gene in the data can mean two things: absence (or frequency too low to measure) of mutations in the gene in the tumor and/or absence of the gene on the panel. In the latter case we simply do not know if there is a mutation or not.

This complication can lead to incorrect conclusions or incorrect computations, for instance of the frequency of a mutation across a cohort or when comparing mutational burden between two samples that have been sequenced on different panels.


The challenges surrounding the usage of different gene panels were already recognized by the cBioPortal community. This led to the implementation of gene panels in the data model and the oncoprint visualization.

A gene panel in cBioPortal is basically a list of genes that are on the panel, stored together with some metadata. This allows us to distinguish between not sequenced and no mutation found in the oncoprint. The computation of the mutation frequency was changed to number of mutated samples for gene x divided by number of samples profiled for gene x, instead of number of mutated samples for gene x divided by all sequences samples.

Below you see an example of the oncoprint with TP53, which is of course on all gene panels, and ZFHX3, which can only be found on the newer MSK IMPACT 400 panel and not on MSK IMPACT 341, as indicated by the dashes.

Gene panels in the patient view

One of the key features in the patient view is the mutations table, where for each mutation the samples are listed that have that particular mutation. The mutations are accompanied by annotations from various sources that show what we know about the mutation, its oncogenicity and, hopefully, potentially effective treatment options. See below for an example from a patient that has 5 samples.

If you look at row three and four, ZFHX3 and FOXA1, you can see that ZFHX3 is not profiled on the first three samples as indicated by the dashes, while FOXA1 was profiled but not found. Before we implemented support for gene panels in the patient view this distinction was not visible, and it would seem that both mutations were not found in the first three samples. We can now also assume that in sample 5, both mutations are not present, instead of first having to check if the genes are both included in the gene panel that was used. This allows the oncologist or researcher to get more insight in how the tumor develops over time without having to figure out which genes were profiled for which sample.

To make it even easier, especially when looking at samples with very different gene panels, for instance when comparing tumor tissue DNA with ctDNA, we have implemented a filtering option for the table. The filter can be set by clicking the filter icon in the ‘Gene’ table header. Setting the option to ‘Genes profiled in all samples’ will leave out all genes that were not included in all panels used for this patient. For this patient it would look like this:

The filtering also influences the genomic overview at the top and the variant allele frequency, see screenshot below, although for this patient that is hardly visible.

We have added labels next to the sample tracks to indicate the panel that was used. Hover over these to see the name of the panel and, if you click the name, the genes that are included on the panel. This is helpful if you want to check that a sample was actually profiled for a gene that was not listed in the table.

Closing remarks

The Hyve provides services to develop, extend and improve features in cBioPortal. Implemented features are released to the community via the cBioPortal repository on GitHub. We hope you like this new feature and it proves useful in your work. For inquiries on cBioPortal feature development projects or other services around cBioPortal, feel free to contact us.

Written by

Sjoerd van Hagen