Hello Community,
we have just published our tool for outlier-aware negative binomial regression. Using an iterative approach for outlier detection.
Now, even more impactful (and trendy) in genetics will be the next tool sccomp, for dirichlet-multinomial (or even better dirichlet negative multinomial) regression and outlier detection, for compositional analysis of single-cell data.
Although the base tool is ready it remains to:
- prove calibration
- benchmark the method against existing ones
- develop possibly the dirichlet negative multinomial regression ( Dirichlet negative multinomial regression for overdispersed correlated count data (nih.gov))
I am happy to coordinate the work, and leave as first author the person who will complete the project. I am also happy to share last authorship with the lab head the student is from.
Please let me know if interested. Spread the word please ;)