The codes here are used to explore if the chromatin loop information, i.e. ChIA-PET data, together with other epigenomics and transcriptomics data could contribute to the transcription and exon usage prediction.
The datasets can be found at: https://researchdata.ntu.edu.sg/dataverse/chrom_pred_exon.
fold10_tran_2.py -- 10-fold cross validation for transcription prediction.
chrom_tran_2.py -- chromsome split validation for transcription prediction.
cross_tran_2.py -- cross cell line validation for transcription prediction.
fold10_exon_coefftreat_2.py -- 10-fold cross validation for exon usage prediction.
chrom_exon_coefftreat_2.py -- chromsome split validation for exon usage prediction.
cross_exon_coefftreat_2.py -- cross cell line validation for exon usage prediction.
The overview of the whole pipeline illustrated in Figure 1.
figure 1.An overview of the pipeline.
Based on python2.
tested on Linux
Python modules:
numpy
pandas
sklearn
If you have any inqueries, please contact mfullwood@ntu.edu.sg.