Hey guys @himel.mallick ,
This tool is so seamless but I have a hard time understanding the boxplot results. I am a master’s student trying to do differential abundance testing but is not well-versed in stats.
My experiment is longitudinal on a synthetic bacterial consortium grown on different carbon sources.
I used Maaslin to check which species in the consortium are differentially abundant in a given carbon source. I got a result with a bunch of boxplots and for a specific species the plots look the same but the “value variable” ( here it is each carbon source)changes with each plot. I want to know why the box plots are constructed multiple times under different “value”.
Above i see that in both blautia plots, the box plots remain the same in pattern but there are two different plots constructed on each “value” variable.
I also set a carbon source condition “MUCIN,HMO,1” yet it somehow draws coefficients in the table. How is the coefficient value so high produced when the species from same as the reference (see below)
It’s hard to tell without seeing the code / input data, but my first guess is that the commas in your CarbonSource variable are messing with the handling of factor levels. Try replacing those with underscores or dashes or something.
Hey @andrewGhazi ,
Thank you so much for your prompt reply. Yes I did change the commas to underscores and saw a vastly different result even eliminating the reference variable from the output table. I still have the doubt on understanding the box plot. What does the plots represent ?
updated plot
The boxplots show the filtered data. For categorical covariates like in your example, the labels in the top right will refer to specific significant associations comparing a specific level (e.g. MUCIN_HMO) against the reference level (which looks to be MUCIN in your case). There will be some boxplots that repeat the same data but have labels describing a different associations for the other significant levels (MUCIN_HMO_1, …).
