<img height="1" width="1" src="https://www.facebook.com/tr?id=1582471781774081&amp;ev=PageView &amp;noscript=1">

3 Tips to Navigate a PDX Database

representative image of in vivo models and data found in a PDX database

representative image of in vivo models and data found in a PDX databaseFind out how to select the right patient-derived xenograft model database for you, and which features to use for quick and easy PDX model selection.

Patient-Derived Xenograft Model Databases

Patient-derived xenograft (PDX) models are an ideal preclinical tool to understand the efficacy of a drug before moving into the clinic. With the increasing number of models available from multiple sources, PDX databases featuring these models and their important characterization data have emerged across the industry.

PDX databases often have hundreds, or even thousands, of models to search. Therefore, it’s important to understand how best to navigate around these useful tools, to make searching for the ideal PDX model as easy as possible.

Identify the Database Which is Best for You

There’s now a wide variety of PDX databases available across the industry, all providing access to enormous amounts of data to enable model selection (I’ll save a review of these different options for a later post). However, there are a few key questions to ask when beginning to search for PDX models. These questions will help narrow down the database which will be most useful to you.

The key questions are:

  • Do you have a specific provider you are looking to use and outsource your study? If so, does that provider offer a database to explore their models in more detail?
  • Are you undecided on a provider and want to compare many models across different institutions from one source?
  • Are you looking for academic institutions you can source PDX models from and perform the studies yourself?
  • Are you looking for a provider with a large selection of models who can perform your studies directly?
  • Are you going to just google “PDX model database” and see what comes up?

Once you answer these questions, you’ll understand where you need to go to begin your search.

Almost all CROs who perform PDX studies offer their own databases to search across available models. These individual company databases contain hundreds or thousands of PDX models to choose from (dependent on collection size). The CRO can usually also perform all the study services you need.

New database tools, such as PDXFinder and Repositive, collate many different PDX model collections from academic labs and some CROs. You’re then able to search across different collections, with these sites helping communication with a specific provider once you’ve identified the models you need.

And finally, Google can help you get to many of these sites just by searching!

Narrow Down the PDX List

The first thing to look for when entering a new PDX database is how easy is it to filter/search and narrow down the list of PDX models you’ll be evaluating. It’s important to be prepared with a list of high-level features critical to your research.

For example, are you looking for a large collection of lung cancer models to choose from, or for a set of models featuring a specific mutation? Do you want to make sure the PDX model has a growth rate which fits your study timelines or has been screened for a specific standard of care?

A good filter system should include:

  • Tumor type.
  • Ethnicity.
  • Growth characteristics available.
  • Standard of care data available.
  • RNAseq data available.
  • Information on the specific drugs the model has been screened with.

Once you’ve filtered your search at this basic level, the number of models for you to review should be narrowed down. You can now focus on more in-depth and specific data to evaluate each PDX model on your list.

Understand the Importance of Genotypic Characterization

With a narrowed-down list of PDX models to analyze, the next step is to understand the breadth of characterization data available. This should have a specific focus on the type of genotypic characterization of each model. Well-characterized models and a good database will typically include several of the following datasets:

  • Gene expression (RNAseq).
  • Copy number (WES).
  • miRNA expression.
  • Mutation.
  • Gene fusion, or combinations of these features.

This type of data is critical when comparing PDX models.

For example, are you looking for models with a high gene expression level of BRAF with a specific copy number variant? This type of detailed data allows you to narrow down your model pool to a very limited number of PDX at this point.

When looking over this genomic data, it’s important to see if the data is easy to export to make your analysis more straightforward. You should also be able to compare gene expression levels across single and multiple genes of interest, with results easily viewed in both graphical and tabular format. I also recommend watching any tutorial videos available to fully understand how best to use these features.


Selecting the right model for your PDX study is an important task and there are many resources at your fingertips to help you find the model which best fits your research. Having a good understanding of how to navigate these databases and the types of characterization data available can point you in the right direction from the start of your search.

Related Posts