Add chemical similarity matrices to the mmo

This function reads cosine, DREAMS, and MS2DeepScore molecular networking outputs from MZmine and stores the raw pairwise similarity matrices (values 0–1, higher = more similar) in the mmo. For datasets with more than 10,000 features a sparse Matrix is used to avoid allocating a full dense matrix; unobserved feature pairs carry implicit similarity 0 (equivalent to dissimilarity 1, maximally different).

Usage

AddChemSim(mmo, cos_dir = NULL, dreams_dir = NULL, m2ds_dir = NULL)

Arguments

mmo

The mmo

cos_dir

Path to the cosine similarity CSV file from MZMine (molecular networking)

dreams_dir

Path to the DREAMS similarity CSV file from MZMine (molecular networking)

m2ds_dir

Path to the MS2DeepScore similarity CSV file from MZMine (molecular networking)

Value

The mmo with similarity matrices stored in mmo$cos.sim, mmo$dreams.sim, and/or mmo$m2ds.sim. Large datasets (n > 10,000) use sparse dgCMatrix format; smaller datasets use a dense numeric matrix.

Note

Tree-based downstream functions (GetFaithPD, FeatureDendrogram, GetBetaDiversity with method = "Gen.Uni") require a full n x n distance matrix and will fail at large scale regardless of sparse storage. Use filter_mmo() to reduce feature count before calling those functions.

Examples

if (FALSE) {
mmo <- AddChemSim(mmo,
 cos_dir = "path/to/cosine_similarity.csv",
 dreams_dir = "path/to/dreams_similarity.csv",
 m2ds_dir = "path/to/ms2deepscore_similarity.csv"
)
}