Laurent Lacroix (laurent.lacroix@inserm.fr)
Raw data are available from ENA repository under accession number PRJEB50302.
A fast5 sample from the BT1_run4 is available from the Zenodo repository under the DOI 10.5281/zenodo.5958270.
In order to test this procedure, the BT1_run4.tar.gz file should be downloaded and expanded into a BT1_run4_fast5 folder.
basecalling_sample.sh contains a example of the base calling procedure going from RawData to the megalodon mod_mappings.bam file.
basecalling_sample.r proceeds in R from the mod_mappings.bam file to smoothed signal in a .rds format (to open with R) used for forks detection. In this script the bam file is first split into 50 000 reads bam files to allow better memory handling. This process requires samtools (we used v1.10). Please note that the script should be changed if more than 100 subfiles are created (experiment with more than 5 millions reads)
Parsing_function4megalodon.r contains the function used in the basecalling process going from the modified bam file created by megalodon to the parsed data used to detect th forks. The parsed data contain both the raw data and the smoothed data (see publication). This script also annotated gaps that can be introduced during the alignment step by megalodon. These gaps will be used during the forks detection process to exclude wrongly detected forks due to a signal discontinuity introduced by those gaps.
session_info() : see base_calling_session_info.txt