In October 2019, we participated in the BioCompute Object (BCO) App-a-thon organized by PrecisionFDA partnered with George Washington University and FDA/CBER HIVE (US Food and Drug Administration High Performance Virtual Environment group). The app-a-thon task was to create a BCO for at least two-step bioinformatics pipeline and/or create an application, which helps with the creation, display, and test of a BCO.
Without standardized schemas for reporting bioinformatics workflows (which often consist of tens of steps, hundreds of parameters, and versions), results’ reproducibility, workflow presentation, and evaluation are very hard. Hence, BCO was created.
BCO (BioCompute Object) is a documentation of the bioinformatics pipeline in JSON format, which contains all the information necessary to understand and replicate the pipeline. Due to its flexibility and openness, it can be adapted to different types of workflows. BCO helps to build transparency in the bioinformatics analysis process and improves reproducibility.
The purpose of the App-a-thon was to encourage the use and development of the BioCompute specifications. Our submitted pipeline consisted of two steps, taking two paired FASTQ files as an input and producing BAM file. The first step consisted of checking the sample quality (using the FastQC tool) and generating reports about it. The second step was to align the readings to the reference genome using BWA-MEM. In both steps, the BCO files were generated with the program we created. The final step was to check the files for compliance with the specification and combined them into a coherent report in PDF, DOCX and ODT formats.
We are proud to announce that our team received the highest scores (gold medals) for our BCO in all three categories: functionality, documentation, and usability/aesthetic. Thanks to the organizers for the recognition!
see the results here: https://precision.fda.gov/challenges/7/view/results