Analysis of co-culture by three methods: univariate analysis (Test of Student or Test of Wilcoxon), OPLS-DA (Orthogonal Projections to Latent Structures Discriminant Analysis) or POCHEMON (Projected Orthogonalized Chemical Encounter MONitoring).

**data** A matrix with the data which will be analyzed, samples in rows, variables in columns.

**SampleNames** A vector of Names corresponding to the different samples.

**monocultureSamples** A vector of Names corresponding to the different monoculture names (should be present in SampleNames).

log10L

**cocultureSamples** A name of the co-culture samples (should be present in SampleNames).

**Method** Different methods which can be used. ‘Ttest’ for Test of Student, ‘WTest’ for Test of Wilcoxon, ‘OPLS-DA’ for the OPLS-DA and ‘Pochemon’ for Pochemon.

SaveImage

**scaleC** Centering and scaling options. ’none’ for either no centering nor scaling, ‘standard’ for mean-centering and unit variance scaling, ‘center’ for mean-centering only and ‘pareto’ for mean-centering and pareto scaling. ‘Ttest’ and ‘WTest’ methods only support 'none' centeringC value.

**log10L** Log10 transformation option. TRUE to transform, FALSE if not.

Lim.Pochemon

**lim.FC** Limit for FoldChange in the case of ‘Ttest’, ‘WTest’ methods.

**lim.pVal** Limit of Test of Student and Test of Wilcoxon’s features selection. (Minimal value : 1, optimal value : 2).

**lim.VIP** Limit of OPLS-DA’features selection. (Minimal value : 1, optimal value : 2)

Lim.FC

**lim.Pochemon** Limit of the Pochemon’s features selection. Proportion of features potentially interesting to keep. Value between 0 and 1 (optimal value between 0.9 et 0.95).

**SaveImage** Saving graphs option. TRUE to save the graphs, FALSE if not. Images are saved in the active directory.

<h1>Value</h1>

Lim.VIP Different methods which can be used. ‘Ttest’ for Test of Student, ‘WTest’ for Test of Wilcoxon, ‘OPLS-DA’ for the OPLS-DA and ‘Pochemon’ for Pochemon.

the function return different tables depending on the method choosen.

Centering and scaling options. ’none’ for either no centering nor scaling, ‘standard’ for mean-centering and unit variance scaling, ‘center’ for mean-centering only and ‘pareto’ for mean-centering and pareto scaling.

**Student test or Wilcoxon test**

* Xtest.FoldChange: Score of the fold change by Wtest or Ttest. Calculated by ‘Co-culture initial value’/’max of the initial value between monocultures’ for each variable.

* Xtest.p-value: Score of the significance of the variables calculated by Test of Student (matched data) or Wilcoxon Test (non matched data).

Log10 transformation option. TRUE to transform, FALSE if not

**OPLS-DA**

* OPLS.DA.VIP1 (Monoculture1 / Co-coculture): Score of VIP for the first monoculture. Used to determinate the limit of interesting features’s selection. Calculated by the function opls from ‘ROPLS’ package.

* OPLS.DA.VIP2 (Monoculture2 / Co-coculture): Score of VIP for the second monoculture. Used to determinate the limit of interesting features’s selection. Calculated by the function opls from ‘ROPLS’ package.

* OPLS.DA.p1 (Monoculture1 / Co-coculture): Score of loadings for the first monoculture. Calculated by the function opls from ‘ROPLS’ package.

* OPLS.DA.p2 (Monoculture2 / Co-coculture): Score of loadings for the second monoculture. Calculated by the function opls from ‘ROPLS’ package.

* OPLS.DA.p1+p2: Sum of OPLS.DA.p1 and OPLS.DA.p2.

Saving graphs option. TRUE to save the graphs, FALSE if not. Images are saved in the active directory.

**Pochemon**

* Pochemon.PC1: Score of loadings for the first axe after singular value decomposition. Calculated by the function ‘svd()’.

* Pochemon.PC2: Score of loadings for the second axe after singular value decomposition. Calculated by the function ‘svd()’.

* Pochemon.SSRankProduct: Rank of interest for each variable after normalization (between 0 and 1).

* Pochemon.Selected: Selected features by Pochemon method.

Limit of the Pochemon’s features selection. Proportion of features potentially interesting to keep. Value between 0 and 1 (optimal value between 0.9 et 0.95)

Limit of Test of Student and Test of Wilcoxon’s features selection. (Minimal value : 1, optimal value : 2)

Limit of OPLS-DA’features selection. (Minimal value : 1, optimal value : 2)

<h1>Author(s)</h1>

Julien WANG, Samuel BERTRAND

<h1>References</h1>

Value

• Test of Student or Test of Wilcoxon

o Xtest.FoldChange: Score of the fold change by Wtest or Ttest. Calculated by ‘Co-culture initial value’/’max of the initial value between monocultures’ for each variable.

o Xtest.p-value: Score of the significance of the variables calculated by Test of Student (matched data) or Wilcoxon Test (non matched data).

• OPLS-DA

o OPLS.DA.VIP1 (Monoculture1 / Co-coculture): Score of VIP for the first monoculture. Used to determinate the limit of interesting features’s selection. Calculated by the function opls from ‘ROPLS’ package.

o OPLS.DA.VIP2 (Monoculture2 / Co-coculture): Score of VIP for the second monoculture. Used to determinate the limit of interesting features’s selection. Calculated by the function opls from ‘ROPLS’ package.

o OPLS.DA.p1 (Monoculture1 / Co-coculture): Score of loadings for the first monoculture. Calculated by the function opls from ‘ROPLS’ package.

o OPLS.DA.p2 (Monoculture2 / Co-coculture): Score of loadings for the second monoculture. Calculated by the function opls from ‘ROPLS’ package.

o OPLS.DA.p1+p2: Sum of OPLS.DA.p1 and OPLS.DA.p2.

• Pochemon

o Pochemon.PC1: Score of loadings for the first axe after singular value decomposition. Calculated by the function ‘svd’.

o Pochemon.PC2: Score of loadings for the second axe after singular value decomposition. Calculated by the function ‘svd’.

o Pochemon.SSRankProduct: Rank of interest for each variable after normalization (between 0 and 1).

o Pochemon.Selected: Selected features by Pochemon method.

Bertand et al., “Metabolite induction via microorganism co-culture: A potential way to enhance chemical diversity for drug discovery”, Biotechnol Adv (2014), http://dx.doi.org/10.1016/j.biotechadv.2014.03.001

Jansen et al., “Projected Orthogonalized Chemical Encounter MONitoring (POCHEMON) for microbial interactions in co-culture”, Metabolomics (2015) 11:908-919, DOI 10.1007/s11306-014-0748-5

Jansen *et al.*, “Projected Orthogonalized Chemical Encounter MONitoring (POCHEMON) for microbial interactions in co-culture”, Metabolomics (**2015**) 11(4):908-919, http://dx.doi.org/10.1007/s11306-014-0748-5

Bertand *et al.*, “Metabolite induction via microorganism co-culture: A potential way to enhance chemical diversity for drug discovery”, Biotechnology Advances (**2014**) 32(6):1180-1204, http://dx.doi.org/10.1016/j.biotechadv.2014.03.001