User story
As a biologist, I want to know how the quality of my samples compares to others, so that I can determine whether or not my results are trust-worthy. If my samples failed, I want to know why.
Describe the solution you'd like
RSeq should implement a model which provides an interpretable 'quality score' based on the features generated by the upstream RSeq pipeline (e.g., "percent of reads aligned" and "exon enrichment"). The model should also be reasonable given machine learning best practices and be defensible to reviewers.
Known steps required
[ ] Research the problem. Is a linear model appropriate here? What other alternatives are there?
[ ] Then, feature selection. Speak to @millerh1 to verify whether these features are feasible.
[ ] Then, feature engineering if needed.
[ ] Then, choose final model and build it.
[ ] Write corresponding methods and update the RSeq report template accordingly.
User story As a biologist, I want to know how the quality of my samples compares to others, so that I can determine whether or not my results are trust-worthy. If my samples failed, I want to know why.
Describe the solution you'd like RSeq should implement a model which provides an interpretable 'quality score' based on the features generated by the upstream RSeq pipeline (e.g., "percent of reads aligned" and "exon enrichment"). The model should also be reasonable given machine learning best practices and be defensible to reviewers.
Known steps required
Depends upon
5