miRbiom: Machine-Learning on Bayesian Causal Nets of RBP-miRNA interactions successfully predicts miRNA profiles

The user needs to provide the RNA-seq or any expression profiling experiment data for any given condition. This data is run through the trained XGBoosting models which generates a relative expression scores for various miRNAs capturing the potential expression profile of the miRNAs for the given condition. For validation purpose, a benchmarking benchmarking script is also provided in the benchmarking directory where the user can provide their actual experimental profiling data for the miRNAs and compare the predicted profile.

The miRNA profile prediction system has been implemented as a webserver at https://scbb.ihbt.res.in/miRbiom-webserver/.

1. Python2.7
2. Numpy
3. Pandas
4. joblib
5. sklearn
6. xgboost
7. Models for all 1,204 miRNAs, download it from https://scbb.ihbt.res.in/miRbiom-webserver/SC/xgb_all.tar.gz (extract it into the parent folder)

1. ensid = ensembl IDs.
2. mirna-list-id = List of mature miRNAs ID, mirbase id and mirTarBase ids.
3. xgb.py = Python script for predicting expression of mature miRNAs.
4. Gene_expression.txt = Example file for user given input[Two column file: First column has ensembl gene ID, second column has expression values for these gene IDs.]
5. example1.txt, example2.txt , example3.txt = Example files for user given input[Two column file: First column has ensembl gene ID, second column has expression values for these gene IDs.]

./benchmarking

1. validation.py = Python script for benchmarking
2. test.csv = Example file for benchmarking [Three column file: First column has miRNAs name, second column has Actual expression or expression came from expreimental data , third column has predicted expression or expression predicted by mirBiom tool.In this file example1.txt predicted and actual expression is used]
3. example1_miRNA_actual_expression.txt , example2_miRNA_actual_expression.txt, example3_miRNA_actual_expression.txt = Actual miRNA expression came from expreimental data for example1.txt , example2.txt , example3.txt respectively.

./xgb_all [Warning: Following directory contain model files, kindly don't try to touch or modify these files]

To predict the expression of mature miRNAs, In parent directory execute following command:

python xgb.py example1.txt > example1_out.txt

To benchmark the output, execute following command in benchmark directory: echo "miRNA,Predicted,Actual" >test.csv awk '{print $1}' example1_out.txt |while read i; do printf "$i,"; grep -P "^$i " example1_out.txt |awk '{printf $3}'; grep -P "^$i\t" example1_miRNA_actual_expression.txt | awk '{printf ","$2}' ; echo "";done |awk -F"," '{if(NF==3) print $0}' >>test.csv

python validation.py test.csv[User defined input]

Mature miRNAs expression prediction module (xgb.py) gives output in following format [ example1_out.txt]:

hsa-let-7a-2-3p[mature miRNAs ID] , 10.608787[Predicted expression value] , MI0000061[mirBase ID] , MIRT058253[mirTarBase ID], logarithm[log2]

Benchmarking script (validation.py) gives output in following format:

Root Mean Square Error (RMSE): 1.0670394705731083 [RMS Error value] Mean Absolute Percentage Error (MAPE): 8.943718334276047 [MAP Error value] Accuracy(%): 91.06 [Accuracy value]

Citation: Pradhan UK, Sharma NK, Kumar P, Kumar A, Gupta S, Shankar R (2021) miRbiom: Machine-learning on Bayesian causal nets of RBP-miRNA interactions successfully predicts miRNA profiles. PLoS ONE 16(10): e0258550. https://doi.org/10.1371/journal.pone.0258550