class: center, middle, title-slide # <code>annoFuse</code> and <code>shinyFuse</code>: ## Annotating, prioritizing, and interactively exploring putative oncogenic RNA fusions ### <table> <tr> <td width="50%"> <a href="https://federicomarini.github.io">Federico Marini</a> </td> <td width="50%"> <a href="https://d3b.center/team-members/gaonkar/">Krutika Gaonkar</a> </td> </tr> <tr> <td align="center" width="50%"> IMBEI, University Medical Center, Mainz </td> <td align="center" width="50%"> Center for Data Driven Discovery in Biomedicine, Philadelphia </td> </tr> <tr> <td align="center" width="50%"> <a href="https://twitter.com/FedeBioinfo"><code><i class="fab fa-twitter "></i> @FedeBioinfo</code></a> </td> <td align="center" width="50%"> <a href="https://twitter.com/29aprilkrutika"><code><i class="fab fa-twitter "></i> @29aprilkrutika</code></a> </td> </tr> </table> ### <p>2020/12/14<br> <a href="https://eurobioc2020.bioconductor.org/"><code>EuroBioc2020</code></a><br><br><a href="https://bit.ly/biocFuse"><code>https://bit.ly/biocFuse</code></a><br></p> --- layout: true <!-- Submitted abstract: --> <!-- Gene fusion events are a significant source of somatic variation across adult and pediatric cancers, and have provided some of the most effective clinically-relevant therapeutic targets. Still, computational algorithms for fusion detection from RNA sequencing data show low overlap of predictions across methods. Additionally, events such as polymerase read-throughs, mis-mapping due to gene homology, and fusions occurring in healthy normal tissue require stringent filtering, making it difficult for researchers and clinicians to discern gene fusions that might be true underlying oncogenic drivers of a tumor and in some cases, appropriate targets for therapy. --> <!-- Here, we present annoFuse and shinyFuse, an R package with a companion Shiny application, developed to annotate, identify, and explore biologically and clinically-relevant expressed gene fusions in single samples, as well as recurrent fusions in a given cohort. --> <!-- We applied annoFuse to STAR-Fusion and Arriba results for 1,028 pediatric brain tumor samples provided as part of the Open Pediatric Brain Tumor Atlas (OpenPBTA) Project. First, we used FusionAnnotator to identify and filter “red flag” fusions found in healthy tissues or in gene homology databases. Using annoFuse, we filtered out fusions known to be artifactual and retained high-quality fusion calls. Second, we prioritized and captured known and putative oncogenic driver fusions previously reported in TCGA or fusions containing gene partners that are known oncogenes, tumor suppressor genes, genes documented in the COSMIC database, or transcription factors. Finally, we enable seamless exploration of the output from the annoFuse workflow by providing a web application, shinyFuse, where users can explore fusion calls in interactive tables augmented with links to external databases, visualize breakpoints by transcript with domain annotation, and plot recurrent fusions within cohorts. --> <!-- Fusions are often associated with early tumorigenic events, therefore we anticipate that systematically prioritizing fusion data with annoFuse will help inform experimental and therapeutic studies. --> <!-- The annoFuse package is available at https://github.com/d3b-center/annoFuse, and we expect to submit it to the Bioconductor project. --> <!-- TL;DR: annoFuse provides standardized filtering and annotation for gene fusion calls, in single samples and in large cohorts, for effectively exploring, prioritizing, and translating fusion findings. --> --- # Gene Fusions <p align="center"> <img src="images/gene_fusion_translocation2.png" height="250"/> </p> .small[https://en.wikipedia.org/wiki/Fusion_gene, CC-BY-SA: Guy Leonard] -- - Frequent source of somatic variation (adult + pediatric cancers) - Among most clinically-effective therapeutic targets - Fusion calls: a variety of algorithms exist for detection from RNA-seq data --- # Gene Fusions <p align="center"> <img src="images/gene_fusion_translocation2.png" height="250"/> </p> .small[https://en.wikipedia.org/wiki/Fusion_gene, CC-BY-SA: Guy Leonard] Yet: - Still a considerable amount of candidates - Low consensus of the predictions - Informed filtering is necessary (polymerase read-throughs, mis-mapping due to gene homology, fusions in normal tissues) -- **It is difficult to pinpoint fusions as true underlying oncogenic drivers!** --- # Introducing `annoFuse`! `annoFuse` - [`https://github.com/d3b-center/annoFuse`](https://github.com/d3b-center/annoFuse) - helps you - **annotating** - **prioritizing** - **exploring** and **visualizing** biologically-relevant gene fusions -- ... on single samples, as well as on large patient cohorts. -- <br><br><br> *A small ode to the Bioconductor community*: this collaboration was born at the virtual poster tables of Bioc2020! <!-- not everything happening in 2020 is bad --> Preprint available! [`https://www.biorxiv.org/content/10.1101/839738v3`](https://www.biorxiv.org/content/10.1101/839738v3) *"annoFuse: an R Package to annotate, prioritize, and interactively explore putative oncogenic RNA fusions"* <!-- we need a sticker soon! --> --- background-image: url("pipeline_steps/Slide1.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow --- background-image: url("pipeline_steps/Slide2.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow <!-- --- --> <!-- # Fusion standardization --> <!-- Input: fusion calls from STAR-FUSION or Arriba (top ranked algorithms in recent benchmarking) --> <!-- Possibility to merge calls from different software --> <!-- Out: tabular tsv data, fundamental for all other calls --> --- background-image: url("pipeline_steps/Slide3.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow <!-- # Fusion filtering --> <!-- - Polymerase read-throughs, mis-mapping due to homology --> <!-- - filter out artifacts, retain high-quality set for later steps --> --- background-image: url("pipeline_steps/Slide4.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow <!-- # Fusion annotation --> <!-- - enhance the information at hand with expression level info for genes involved in fusion --> <!-- - adds orthogonal info from other databases, extremely useful for prioritization: TCGA, known as oncogenes, as tumor suppressor, known in the COSMIC database, or as transcription factors --> <!-- - on top of that, domain annotation via the Pfam database is also super handy to inspect how the functionality of those proteins might be affected --> --- background-image: url("pipeline_steps/Slide5.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow <!-- Project-specific --> <!-- - Sometimes you just don't have a handful of samples, so you want to have more refined control on the info over your entire patient cohort --> <!-- - this works quite nicely as we show in our manuscript with a thousand-patient cohort of pediatric brain tumors, for the OpenPBTA use case --> --- background-image: url("pipeline_steps/Slide6.png") background-size: contain background-position: 40% 40% class: center, bottom # The `annoFuse` workflow <!-- # Viz & exploration --> <!-- - last but not least, visualization and exploration, in an interactive way. Why? --> <!-- - because you might still have a considerable amount of candidates which might be hard to digest easily, and you don't want your bottleneck to be data interpretation, which is quite compelling and especially the one where the decisions take place --> --- # Fusion visualizations Running `plot_summary` on the OpenPBTA set... <p align="center"> <img src="images/ss_af_summary_p1.png" height="450"/> </p> --- # Fusion visualizations Running `plot_summary` on the OpenPBTA set... <p align="center"> <img src="images/ss_af_summary_p2.png" height="450"/> </p> --- # Fusion visualizations Running `plot_recurrent_genes` and `plot_recurrent_fusions` <p align="center"> <img src="images/ss_af_recurrent_fusions.png" height="450"/> </p> --- # Fusion visualizations Running `plot_breakpoints` on a single fusion <p align="center"> <img src="images/ss_af_domain.png" height="350"/> </p> --- background-image: url("images/ss_af_shinyfuse.png") background-size: contain background-position: 50% 50% class: center # Exploring with `shinyFuse` <br><br><br><br><br><br><br><br><br><br><br><br><br><br><br><br><br><br><br> Demo instance: [`http://shiny.imbei.uni-mainz.de:3838/shinyFuse/`](http://shiny.imbei.uni-mainz.de:3838/shinyFuse/) --- # Summary - Road to Bioconductor `annoFuse` can annotate, prioritize, and explore/visualize putative oncogenic RNA fusions -- - Fully functional package, with all workflow operations - Vignettes describe complete use cases to showcase functionality - Rich set of visualizations to generate insights - Interactivity and reproducibility: `shinyFuse` and `reportFuse` to deliver a combination of flavours -- - Best strategy to provide mid-size datasets, required for functionality - Unit tests -- **Outlook**: extend tools to support, add gene expression related functionality, domain retention linked to drug database to inform targeting strategies <!-- Evaluation + translation of findings in clinical settings! --> [`https://github.com/d3b-center/annoFuse`](https://github.com/d3b-center/annoFuse) [`http://shiny.imbei.uni-mainz.de:3838/shinyFuse/`](http://shiny.imbei.uni-mainz.de:3838/shinyFuse/) --- class: center ## ... thank you for your attention! You can find us here: <code><i class="fas fa-envelope "></i> marinif@uni-mainz.de - <i class="fab fa-twitter "></i></code> [`@FedeBioinfo`](https://twitter.com/FedeBioinfo) <code><i class="fas fa-envelope "></i> gaonkark@email.chop.edu - <i class="fab fa-twitter "></i></code> [`@29aprilkrutika`](https://twitter.com/29aprilkrutika) <code><i class="fas fa-desktop "></i> </code>[`https://bit.ly/biocFuse`](https://bit.ly/biocFuse) -- ### Acknowledgements .pull-left[ <i class="fas fa-file-code "></i> <i class="fas fa-lightbulb "></i> **Krutika Gaonkar** <br> Komal Rathi <br> Payal Jain <br> Yuankun Zhu <br> Nicholas Chimicles <br> Miguel Brown <br> Ammar Naqvi <br> ] .pull-right[ Bo Zhang <br> Phillip Storm <br> John Maris <br> Pichai Raman <br> Adam Resnick <br> <i class="fas fa-university "></i> **Konstantin Strauch** <br> Jaclyn Taroni <br> <i class="fas fa-lightbulb "></i> <i class="fas fa-university "></i> <i class="fas fa-file-code "></i> **Jo Lynne Rokita** <br> ] </br> IMBEI - Genomic Statistics & Bioinformatics division, UMC Mainz Center for Thrombosis and Hemostasis (CTH), Mainz (Virchow Fellowship) Center for Data Driven Discovery in Biomedicine (D3b, Philadelphia)