Environment Requirement:
Please visit https://github.com/picrust/picrust2/blob/master/INSTALL.md to install PICRUSt2 with conda and activate the conda environment before running the pipelines.


Input files:
All the raw OTU tables are in folder "mapsseq-master/otu_tables/". The sample information is in file "dataset_info.xlsx".


Pipeline:
1. Run bimapseq.m in matlab to get the basis matrix and coefficient matrix, which will be saved in folder "outputs/". Two files will be saved for each sample, {sample_name}.basis.mat.txt is the basis matrix and {sample_name}.coef.mat.txt is the coefficient matrix. (Matlab R2020a is recommended)

2. Run prepare.for.picrusst.R in R to prepare for the standard input for PICRUSt2. This will create a folder "picrust2-2.3.0-b/mapseq.input/" and save all the required files in it.

3. Run run.picrust.sh in bash shell to perform the PICRUSt2, the output files will be saved under "picrust2-2.3.0-b/{sample_name}.out/"

4. To reproduce the figures in the paper, run draw.bernoulli.mf.mxLH.R, top10.metagenome.R and top10.pathways.R. The figures of bernoulli MF, basis mat and coef mat  will be saved in folder "bernoulli.mf.plots/{sample_name}/". The figures of pathway plot will be saved in folder "pathway.plots/{sample_name}/".