The script used for analyses in "Novel metric for hyperbolic phylogenetic tree embeddings"

Reference

Requirements

The following libraries are used.

• hydra
• MASS
• ape
• ROCR
• RSpectra

Full node embeddings

Rscript embedding_Dall.R

The MSE values of each embedding method for the i-th simulation tree are saved in "out/MSE_Dall_1.csv". The mean MSE values are saved in "out/MSE_Dall_all.csv".

External node embeddings and folding-in internal nodes

Rscript embedding_Dall.R

The MSE values of each embedding method for the i-th simulation tree are saved in "out/MSE_Dint_i.csv". The mean MSE values are saved in "out/MSE_Dint_all.csv".

External node embeddings and folding-in internal nodes of the partial tree

Rscript embedding_Dall.R

The MSE values of each embedding method for the i-th simulation tree are saved in "out/MSE_Dpart_i.csv". The mean MSE values are saved in "out/MSE_Dpart_all.csv".

Prediction of the MRCA

Rscript predicting_MRCA.R

The AUC values of each embedding method for the i-th simulation tree are saved in "out/AUC_MRCA_i.csv". The mean AUC values are saved in "out/AUC_MRCA_all.csv".

Prediction of nearst-neighbor nodes in the partial tree

Rscript predicting_nearest_neighbor.R

The rank of each embedding method for the i-th simulation tree are saved in "out/Result_NN_MDS_i.csv ", "out/Result_NN_H1_i.csv", and "out/Result_NN_H2_i.csv", respectively. The mean rank are saved in "out/Result_NN_MDS_all.csv ", "out/Result_NN_H1_all.csv", and "out/Result_NN_H2_all.csv", respectively.

Test code for imputation of missing distances

Rscript test_imputation_H1.R 80
Rscript test_imputation_H2.R 80

Two partial trees are derived from "data_imputation/tree_1.txt". These trees contain external nodes "data_imputation/idx1_L80.txt" and "data_imputation/idx2_L80.txt", respectively. 80 external nodes (+ outgroup node) are included in both of the partial trees. The imputed distance matrix are saved in "out_imputation/optD_L80_H1.txt" and "out_imputation/optD_L80_H2.txt", respectively.