Hi, after using the maturity index plugin in qiime2-2018.6 and obtaining the feature importance table, I was wondering if it is possible to extract information regarding how each feature affects the maturity score? I.e. is the top feature increased or decreased in relative abundance as the microbiota matures? Thanks in advance!
Yes; in 2018.6 the feature importance scores are listed (and downloadable) at the bottom of the visualization. In 2018.8 they are produced in a separate artifact.
See the heatmap/clustermap. In 2018.6 this is the 4th plot in the visualization. In 2018.8 this is produced in a separate visualization. In 2018.6, the plot dimensions are fixed and so the feature IDs on the x-axis may not show all IDs, making this difficult to interpret if you are looking for specific features. Use 2018.8 for better granularity (since plot dimensions are not fixed).
In case you do update to 2018.8, note that in 2018.8
maturity-index moved over to q2-longitudinal.
I hope that helps!
Thanks for your help! With the latest maturity index in qiime2-2018.8, it is possible to get MAZ scores for all samples, or only samples that were not used for training the classifier? (i.e. not just the test samples?)
MAZ scores are only calculated on the test samples — so that’s everything but the relatively small subset of control samples that will be used for training.
I had considered putting this pipeline together as a nested cross-validated method to report MAZ on every sample (it’s still in the back of my mind), which would be an improvement over the original implementation of this method, but it gets very complicated. If you or someone you know is interested in implementing this more complicated version, let me know, I’ve already brainstormed some ways to work it out
This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.