2024-06-05 Community Update

[Jen-OMalley]

Hello everyone. Here are the latest OpenScPCA updates! Please feel free to reply here with any questions or comments.

Analyses in-progress

#292 Annotate cell types in Ewing sarcoma samples on the Portal

Since our last update:

• We created a small workflow to annotate tumor cells in Ewing sarcoma samples using multiple methods, including manual annotation with marker genes, CellAssign, CopyKAT, and InferCNV.
• This workflow was run on a subset of samples in SCPCP000015 to obtain tumor cell annotations using each method independently.

Next, we are:

• Starting to validate and finalize the tumor cell annotations obtained from running the workflow. This includes looking at the consensus between each of the CNV methods (CopyKAT, InferCNV) and marker gene expression.
• Validating tumor cell annotations by confirming high expression of EWS-FLI1 targets, the fusion gene present in Ewing sarcoma tumor cells. To do this we are calculating a gene set score for commonly used EWS-FLI1 gene set signatures available in MSigDB.

Coming soon: Look out for a discussion post that will go over the current results of this analysis. We will also be seeking your feedback! This post will be shared in the Analysis category within the next couple of weeks.

Read #292 for more details about this analysis, the scientific goals, planned methods and more.

#364 Doublet Detection

Since our last update:

• We started benchmarking different doublet detection methods across multiple benchmarking datasets. Here we were able to use ground truth annotations of doublets to determine performance of each method on each dataset.

Next, we are:

• Comparing performance of doublet detection methodologies across multiple benchmarking datasets.

Read #364 for more details about this analysis, the scientific goals, planned methods and more.

View all current analysis ideas, and share your own ideas, in the Propose a new analysis category!

What the Data Lab is working on

• The Data Lab team is holding a training workshop from June 10-14 and may be slower to respond to messages.
• We recently conducted information architecture (IA) validation to evaluate how useful and discoverable the OpenScPCA documentation is.
  • Based on the feedback we received, we are adding a new section in our documentation about ensuring reproducibility. Here we will introduce the different strategies used in OpenScPCA that ensure project reproducibility over time.
  • As part of ensuring reproducibility, we will include a GitHub action to run each module in their respective docker containers. This ensures that the module can be run successfully on test data.

Do you need assistance getting set up, have an idea for an analysis that you are unsure about, or just have questions about where to begin with OpenScPCA? We can help! We are happy to provide guidance for getting started and we can even meet one-on-one to chat. Post in the Questions category, email us at scpca@ccdatalab.org, or send a message on our #open-scpca Slack channel.