Reapplication of Enterosignatures

These results were generated by the Reapply Enterosignatures tool available at
- https://enterosignatures.quadram.ac.uk/
- https://enterosignatures.streamlit.app
- https://github.com/apduncan/enterosig_sl

The abundance table provided was fitted to the five Enterosignature model 
described in C. Frioux et al. 2023 (doi:10.1016/j.chom.2023.05.024.)
To do this, taxa names were matched between the provided abundance table (X)
and the five Enterosignature W matrix, which gives the weight of each taxon in 
each signature. The W matrix and X have been been modified so taxa names and 
orders in each match.

For details on this process, see the code in the GitHub repository linked 
above. The important point to note is that the W matrix, and the abundance 
table, provided in this archive are the modified ones.

The files contained are
- w.tsv: 5 Enterosignature model W matrix modified to match input taxa
- h.tsv: Enterosignature weights for each sample in input abundance table
- w_norm.tsv: W matrix total-sum-scaled by signature, so the weights of each
            signature sum to 1.
            Note that using scaled version of the matrices means
            w_norm x h_norm is no longer approximately equal to the input.
- h_norm.tsv: H matrix total-sum-scaled by sample, so the weights of each
            sample sum to 1. This can be interpreted as the relative abundance
            of each ES in a sample.
            Note that using scaled versions of the matrices means
            w_norm x h_norm is no longer approximately equal to the input.
- abundance.tsv: Input abundance matrix matched up to taxa in w.tsv
- model_fit.tsv: Fit between values in each sample in abundance.tsv and the 
            product w x h. This is the cosine similarity between the two 
            vectors.
- taxon_mapping.tsv: Which taxa were considered equivalent between the input 
            abundance and the 5 Enterosignature W model.