Welcome to Miqa! This article will show you how to get started with bioinformatic software testing using Miqa.
Overview
Miqa is an end-to-end software test automation platform where bioinformatics team can easily test and manage their QA processes in one place. You can use our built-in tools to quickly set-up a software testing workflow. Once you are comfortable with Miqa, there are endless possibilities for customization to suit your specific applications and QA needs.
Here, we will cover some basics to onboard you to Miqa 🚀
A. Invite your team
If you're the organization admin on Miqa, you will want to also onboard your teammates to collaborate or keep them in the loop. Follow the instructions on this article to add your teammates as new users to your Miqa instance.
B. Prepare your test data
Test datasets are a crucial component to the evaluation of your bioinformatic software or pipelines. In Miqa, a test dataset is composed of:
Descriptive fields, such as Name, source characteristics (e.g. tissue type, chemistry/assay version), and other fields that will either be useful to categorize the datasets for reporting or to use as variables in your pipeline's CLI/execution
Input files, which can be used for cloud execution of your pipeline on this dataset, as well as for Miqa's other automated tasks such as generating new samples by simulation (downsampling, mutating, or mixing datasets) or metaprocessing (evaluating the input files to generate summary statistics that can be used in reporting).
Selecting Test Data
Selecting Test Data
You may use a different test dataset depending on the type, frequency or scale of your test(s). For example, a downsampled BAM file may be used in a quick smoke test while a computationally complex file may be used in a stress test of your software. Similarly, while your daily smoke tests may be run on one or a handful of datasets, you may choose to test accuracy across hundreds of datasets simultaneously when performing release-level or regression testing.
Below are two articles to help you prepare high-quality test data:
Uploading Test Data
Uploading Test Data
Once your test datasets are ready, you can upload them to Miqa using the following methods - see the step-by-step instructions here:
Direct file upload
From cloud storage
This test data serves as the input of your bioinformatic tools which will be executed in test runs either manually or automatically (using automation triggers).
🚀 Quick-Start Tip
If you're integrating a new pipeline, you will likely want to start from one of the basic dataset-types (no-files or directory-based) to quickly get started, and refine the details of your data model after onboarding. The tutorials below show you how to easily add new test data for this purpose.
C. Generate output files for evaluation
To perform testing and assessment of your bioinformatic pipeline results, Miqa evaluates the files outputted by Executions of particular pipeline versions on your Test Datasets. Output files can be generated by:
Executing your pipeline on the cloud using our Cloud Runner
Executing your pipeline locally or remotely using Custom Integrations
Executing your pipeline locally or remotely and uploading results directly using Offline-Versions
🚀 Quick-Start Tip
During onboarding and integration, start with Offline-Versions run on a dataset as created above, to fast track your testing and reporting and to form your testing strategy ahead of automation.
D. Evaluate QA results from your output files
From your bioinformatic outputs, you can design data tests for reproducible and standardized testing, or compare results from multiple test runs in ad hoc reports, all using a variety of built-in evaluation metrics such as accuracy and concordance, or custom statistics derived from your files. Possible sources of reportable and testable metrics are shown in the diagram below.
Miqa's Reporting module allows you to interactively explore results and summarize QA trends across versions or samples. You can also choose from our library of built-in metrics/statistics or customize your own through JSON configuration. Don't hesitate to reach out to the Magna Labs team if you need a hand!
Check out the Reporting docs collection to get started with your QA analysis.
🚀 Quick-Start Tip
To easily begin building tests for your pipeline, use the Output Explorer tool on the outputs of the Offline-Version uploads generated above. If you're at the onboarding phase, your Onboarding Lead will also help you generate your first data tests.
E. Execute and Automate bioinformatic pipeline tests
Automated Execution
Now, to take your QA up a notch, integrate your pipelines into Miqa for execution via Cloud Runner or Custom Integration so you can execute test runs on them, generate test outputs and examine QA results all in one place.
Here are a few concepts that define the execution layer in Miqa:
A pipeline represents the core execution process that takes in an input and generates an output (or results) that can be evaluated.
A workflow represents an implementation of a pipeline. You may have multiple workflows (or variations) of a given pipeline where the input and output types are the same, but the components of the pipeline might differ.
A workflow variant represents a workflow with a distinct parameter set.
You will work with your Onboarding Lead to define the execution configuration as part of Phase 2 Onboarding.
Automated Testing
Miqa provides you with a few options to execute your pipeline for testing: ad hoc execution, automatic execution through CI (when set-up with automation trigger from continuous integration tools like GitHub, BitBucket, CircleCI or Jenkins), or scheduled execution for nightly or weekly tests.
This Executing a Pipeline article gives you a simple tutorial and video on how to execute your bioinformatic pipeline through Miqa, and Automating Tests & Executions gives an overview of how to automate these executions.
Read to try out Miqa on your own? Feel free to contact our support team at any time through the chat bot directly in Miqa or at the bottom right of this page, or email us at [email protected].
You can also find additional articles and guides in our Doc Library.