"The assessment of similarity vectors of fingerprint and UMLS in adverse drug reaction prediction"
*Explain Data & Codes*

Delta_1 Dataset : 
1-train_data_delta1.mat : Training dataset was obtained from the concatenate of drug similarity vectors and adverse reaction similarity vectors.
2-train_label_delta1.mat : Training dataset label (1: Drug-adverse reaction && 0 : Drug-Indication).
3-test_data_delta1.mat : Test dataset was obtained from the concatenate of drug similarity vectors and adverse reaction similarity vectors.
4-test_label_delta1.mat : Test dataset label (1: Drug-adverse reaction && 0 : Drug-Indication).

* The Delta_2 and Delta_3 datasets are similar to the Delta_1 datasets *

1)DRUG_CID_CODE_1430.mat : The CID code of all 1430 Drug (In the Delta_1, Delta_2 and Delta_3 datasets have been used) is given.
2)Drug_INDICATION_CID_CODE_30835.mat : List of drugs that are each associated with an indication (Based on Sider).
3)DRUG_SIDEEFFECT_ASSOCIATION_1430_5868.mat : Drug-adverse reaction association.
4)Fingerprint_Drug_1430.mat : Fingerprint of all drugs.
5)Indication_CUI_CODE_30835.mat : The CUI code of all Indication in Sider.
6)Sideeffect_CUI_CODE_5868.mat : The CUI code of all adverse reactions in Sider.
7)UMLS_SIMILARITY_SIDEEFFECT_5868 : UMLS similarity between all adverse reactions.


--------------------------------------  --------------------------------------  --------------------------------------
MF data : 

*CS_DATA* -> (A. Poleksic and L. Xie, “Predicting serious rare adverse reactions of novel chemicals,” Bioinformatics, vol. 34, no. 16, pp. 2835–2842, Aug. 2018, doi: 10.1093/BIOINFORMATICS/BTY193)

1)ADR_PHENOTYPE_POLEKSIC.mat : List of collections of Poleksic data ADRs names, whose phenotype information is available in the CTD database. 
2)ADRNAMEPOLEKSIC5868.mat : The CUI code of all adverse reactions in Poleksic Data.
3)DATA_CS_MODEL.mat : Input MODEL:
		3-1. FEATUREVECTORS : Fingerprint of drugs.
		3-2. M : Similarity between drugs.
		3-3. N_PHENOTYPE : Similarity between adverse reactions based on Phenotype.
		3-4. N1 : Similarity between adverse reactions based on UMLS.
		3-5. R : Drug-adverse reactions associations.

4)pheno_GOID_Poleksic.mat : ADR-phenotype pairs. 
5)SE_PHENOTYPE_ASSOCIATION_POLEKSIC.mat : ADR-Phenotype association matrix.
6)Unique_PHENOTYPE_POLEKSIC.mat : List of phenotypes associated with Poleksic data set ADRs.

--------------------------------------
*TMF_DATA* -> (G. X, Z. W, Y. Y, D. Y, T. J, and G. F, “A Novel Triple Matrix Factorization Method for Detecting Drug-Side Effect Association Based on Kernel Target Alignment,” Biomed Res. Int., vol. 2020, 2020, doi: 10.1155/2020/4675395.)

1)ADR_NAME_HAS_PHENOTYPE.mat : List of collections of Mizutani data ADRs names, whose phenotype information is available in the CTD database. 
2)ADR_NAME_MIZUTANI.mat : The CUI code of all adverse reactions in Mizutani Data.
3)Drug_CID_MIZUTANI.mat : The CID code of all Drugs in Mizutani Data.
4)Mizutani_dataset.mat : Input MODEL:
	4-1. chemical : Fingerprint of drugs.
	4-2. side_effect : Drug-sideeffect association matrix.
	4-3. Targets : Drug-target association matrix.

5)Phenotype_SE_ASSOciation_ID.mat : ADR-phenotype pairs. 
6)SE_PHENO_ASSOCIATION_MIZUTANI.mat : ADR-Phenotype association matrix.
7)Target_name_mizutani.mat : List of targets associated with Mizutani data set Drugs.
8)Unique_Phenotype.mat : List of phenotypes associated with Mizutani data set ADRs.


ML_MODEL (code_delta1_NN,code_delta1_RF,code_delta2_NN,code_delta2_RF,code_delta3_NN,code_delta3_RF) :
RF MODEL CODE :  ...\Machine_learning_models\codes\F_RF_code
NN MODEL CODE :  ...\Machine_learning_models\codes\F_NN_code

RUN_ML_MODEL : 

code_delta1_NN.m & code_delta1_RF : F_RF and F_NN on DELTA_1.

code_delta2_NN.m & code_delta2_RF : F_RF and F_NN on DELTA_2.

code_delta3_NN.m & code_delta3_RF : F_RF and F_NN on DELTA_3.

Required data (...\Machine_learning_models\datasets) : 

1)DRUG_CID_CODE_1430.mat
2)Drug_INDICATION_CID_CODE_30835.mat
3)DRUG_SIDEEFFECT_ASSOCIATION_1430_5868.mat
4)Fingerprint_Drug_1430.mat
5)Indication_CUI_CODE_30835.mat
6)Sideeffect_CUI_CODE_5868.mat
7)UMLS_SIMILARITY_SIDEEFFECT_5868.mat

***************************

MF_MODEL (CS_MODEL,TMF_MODEL) :

CS MODEL CODE :  ...\Matrix_factorization_models\codes\CS_code
TMF MODEL CODE :  ...\Matrix_factorization_models\codes\TMF_code

RUN_CS_MODEL : 

CS_ORG.m : The original model
CS_PHEN.m : Add the ADR phenotype feature to the original model 

Required data for CS_MODEL (...\Matrix_factorization_models\datasets\CS_DATA) : 

1)DATA_CS_MODEL.mat

--------------------

RUN_TMF_MODEL : 

tmf_ORG.m : The original model
tmf_PHEN.m : Add the ADR phenotype feature to the original model 
tmf_TARG.m : Add the DRUG TARGET feature to the original model 
tmf_PHEN_TARG.m : Add the DRUG TARGET & ADR phenotype features to the original model.

Required data for TMF_MODEL (...\Matrix_factorization_models\datasets\TMF_DATA) : 

1)Mizutani_dataset.mat

___________________________________________________________________

**Contact us : f.zare@aut.ac.ir  & milad1besharati@aut.ac.ir **