I propose a crucial enhancement to ALDEx2's glm function to support using taxa as explanatory variables and non-taxa variables as outcome variables, catering to the substantial interest within the research community in studying the relationship between the microbiota and diseases. Particularly, integrating logistic regression would be beneficial, as outcomes in these studies are often binomial (disease vs. control), which necessitates a model that can handle binary outcomes effectively, e.g. allowing to set family = "binomial" in the glm function). Currently, ALDEx2's glm function only supports using taxa as an outcome variable and only asses variables that may explain this outcome. The proposed feature would allow researchers to use microbial compositions to predict/explain health conditions, thereby expanding ALDEx2’s utility in microbiome research and aiding in the exploration of how microbiota may contribute to or indicate disease states.
I propose a crucial enhancement to ALDEx2's
glmfunction to support using taxa as explanatory variables and non-taxa variables as outcome variables, catering to the substantial interest within the research community in studying the relationship between the microbiota and diseases. Particularly, integrating logistic regression would be beneficial, as outcomes in these studies are often binomial (disease vs. control), which necessitates a model that can handle binary outcomes effectively, e.g. allowing to setfamily = "binomial"in theglmfunction). Currently, ALDEx2'sglmfunction only supports using taxa as an outcome variable and only asses variables that may explain this outcome. The proposed feature would allow researchers to use microbial compositions to predict/explain health conditions, thereby expanding ALDEx2’s utility in microbiome research and aiding in the exploration of how microbiota may contribute to or indicate disease states.