miRBench 0.1.0

A collection of datasets and predictors for benchmarking miRNA target site prediction algorithms

miRNA target site prediction Benchmarks

Installation

pip install git+https://github.com/katarinagresova/miRBench.git

Examples

Get all available datasets

import miRBench

miRBench.dataset.list_datasets()

['AGO2_CLASH_Hejret2023',
 'AGO2_eCLIP_Klimentova2022',
 'AGO2_eCLIP_Manakov2022']

Not all datasets are available with all splits and ratios. To get available splits and ratios, use the full option.

miRBench.dataset.list_datasets(full=True)

{'AGO2_CLASH_Hejret2023': {'splits': {
      'train': {'ratios': ['10']},
      'test': {'ratios': ['1', '10', '100']}}},
 'AGO2_eCLIP_Klimentova2022': {'splits': {
      'test': {'ratios': ['1', '10', '100']}}},
 'AGO2_eCLIP_Manakov2022': {'splits': {
      'train': {'ratios': ['1', '10', '100']},
      'test': {'ratios': ['1', '10', '100']}}}
}

Get dataset

dataset_name = "AGO2_CLASH_Hejret2023"
df = miRBench.dataset.get_dataset(dataset_name, split="test", ratio="1")
df.head()

	noncodingRNA	gene	label
0	TCCGAGCCTGGGTCTCCCTCTT	GGGTTTAGGGAAGGAGGTTCGGAGACAGGGAGCCAAGGCCTCTGTC...	1
1	TGCGGGGCTAGGGCTAACAGCA	GCTTCCCAAGTTAGGTTAGTGATGTGAAATGCTCCTGTCCCTGGCC...	1
2	CCCACTGCCCCAGGTGCTGCTGG	TCTTTCCAAAATTGTCCAGCAGCTTGAATGAGGCAGTGACAATTCT...	1
3	TGAGGGGCAGAGAGCGAGACTTT	CAGAACTGGGATTCAAGCGAGGTCTGGCCCCTCAGTCTGTGGCTTT...	1
4	CAAAGTGCTGTTCGTGCAGGTAG	TTTTTTCCCTTAGGACTCTGCACTTTATAGAATGTTGTAAAACAGA...	1

Data will be downloaded to $HOME / ".miRBench" / "datasets" directory, under separate subdirectories for each dataset.

Get all available tools

miRBench.predictor.list_predictors()

['CnnMirTarget_Zheng2020',
 'RNACofold',
 'miRNA_CNN_Hejret2023',
 'miRBind_Klimentova2022',
 'TargetNet_Min2021',
 'Seed8mer',
 'Seed7mer',
 'Seed6mer',
 'Seed6merBulgeOrMismatch',
 'TargetScanCnn_McGeary2019',
 'InteractionAwareModel_Yang2024']

Encode dataset

tool = 'miRBind_Klimentova2022'
encoder = miRBench.encoders.get_encoder(tool)

input = encoder(df)

Get predictions

predictor = miRBench.predictors.get_predictor(tool)

predictions = predictor(input)
predictions[:10]

array([0.6899161 , 0.15220629, 0.07301956, 0.43757868, 0.34360734,
       0.20519172, 0.0955029 , 0.79298246, 0.14150576, 0.05329492],
      dtype=float32)

Benchmark all tools on all datasets

python benchmark_all.py OUTPUT_FOLDER_PATH

The script will run all tools on all datasets and will produce a file with suffix _predictions.tsv for each dataset. Predictions from every tool will be saved in separate columns.