Identification of methylation for Nanopore DNA sequencing.
Poreformer is a computational tool for detecting DNA 5mC、4mC and 6mA methylation from Oxford Nanopore reads. It uses a Transfomer model to predict per-read and per-site 5mC、4mC and 6mA methylations and produces a methylation file. Poreformer can call methylation from FAST5 files basecalled with Guppy and provides models for R9.4.1 flowcells.
Please refer to Installation for how to install Poreformer.
- To call modifications, the raw fast5 files should be basecalled by Guppy.
ont-guppy/bin/guppy_basecaller -i ${INPUT_DIR}/BA_NAT -s ${INPUT_DIR}/BA_NAT_guppy -c ont-guppy/data/dna_r9.4.1_450bps_hac.cfg -x cuda:all:100% -r --fast5_out
- Extract fastq and signal information from fast5 file and Align reads using minimap2 and then sort and index the BAM file
./align_index.sh -ref Bacillus_amyloliquefaciens.fa -fast5 ${INPUT_DIR}/BA_NAT_guppy/.fast5 -ref_rev Bacillus_amyloliquefaciens_rev.fa
- Extract features
sh extra_feature.sh -ref Bacillus_amyloliquefaciens.fa -forward_current all_zheng_fin_sort_fin.txt -ref_rev Bacillus_amyloliquefaciens_rev.fa -reversed_current all_fan_fin_sort_fin.txt
sh Poreformer.sh -meth_type mC -feature all_mC_reversed_current_mean_kmeans_6_5.txt
sh Poreformer.sh -meth_type 6mA -feature all_6mA_reversed_current_mean_kmeans_7_6.txt
Please refer to Usage.md for details on how to use Poreformer.