• linux-64 v1.3.2
• osx-64 v1.3.2

conda install

If you find a bug, please add it to the issues on GitHub with a detailed description.

1. Installation via Conda
2. Description
3. Dependencies
4. Workflow
5. Usage
6. Output File Description
7. Error Codes Explanation

To install Magnipore we recommend to use Conda: Magnipore is available for linux-64 and osx-64.

conda install mamba
mamba create -n magnipore -c jannessp magnipore
conda activate magnipore

If you want to basecall your ONT data you also need a Guppy version from Oxford Nanopore Technologies.

Magnipore is a tool written in python3 to analyze and pair-wise compare sequencing samples from Oxford Nanopore Technologies (ONT) sequencing.

Magnipore compares two ONT samples on a signal level to find differential signals between them in single base resolution. Such differences are caused by mutations or modifications. Magnipore classifies these differences and provides the user with a position-wise comparison.

Magnipore depends on/requires other tools to preprocess and analyze the data.

Conda Dependencies - python (>=3.8,<3.11) - h5py>=3.7 - biopython>=1.80 - mafft>=7.508 - matplotlib>=3.6.2 - numpy>=1.23 - scipy>=1.9 - winnowmap>=2.0 - pandas>=1.5 - seaborn>=0.12 - psutil>=5.9 - hdf5plugin>=3.3.1 - ontvbzhdf_plugin>=1.0.1 - pytest>=7.1 - gzip>=1.12 - read5>=1.1.6 - f5c>=1.2 - read5>=1.2.0

For each sample in the comparison, Magnipore takes: - (FASTA) exactly ONE reference sequence - (FAST5) the raw sequencing data from ONT - (optinal FASTQ) optionally basecalls, if you do not have the guppy binary or do not want to basecall the raw ONT data (again).

• Magnipore file (TSV)
  • all compared positions
  • classified into mutation and potential modification
  • with the TD score
  • with the Kullback-Leibler divergence
  • with a bayesian p-Value
• reference sequence alignment file
• stockholm file (significant positions are marked)
• multiple plots about the data of the samples like

If you are not using the conda package replace "magnipore" by "python3 magnipore.py".

magnipore raw_data_first_sample reference_first_sample label_first_sample raw_data_sec_sample reference_sec_sample label_sec_sample working_dir --basecalls_first_sample basecalls_first_sample --basecalls_sec_sample basecalls_sec_sample

magnipore raw_data_first_sample reference_first_sample label_first_sample raw_data_sec_sample reference_sec_sample label_sec_sample working_dir --guppy_bin PATH --guppy_model PATH

• basecallsfirstsample/basecallssecsample containing the demultiplexed FASTQs
• label_first_sample.fastq contains only those reads of the first condition
• label_sec_sample.fastq contains only those reads of the second condition
• be sure that the sequencing_summary.txt is next to your FASTQ files, otherwise provide them using
• -s1, --sequencingsummaryfirst_sample
• -s2, --sequencingsummarysec_sample

magnipore --basecalls_first_sample basecalls_first_sample --basecalls_sec_sample basecalls_sec_sample raw_data_first_sample reference_first_sample label_first_sample raw_data_sec_sample reference_sec_sample label_sec_sample working_dir

Using the same reference sequence for both samples results in no reported mutations. Magnipore will only report potential modifications in this case. If you assume there are mutations between the samples, try to provide different reference sequences containing these mutations.

Complete help messages can be found here!

use either the basecalling arguments or provide basecalls - basecalling arguments: - guppybin : Path to guppy binary - guppymodel : Path to guppy model used for basecalling - (optional) guppydevice : Device used for basecalling (cpu or gpu cuda:0) - provided basecalls (FASTQ) - basecallsfirstsample : Path - basecallssec_sample : Path

The .magnipore file is a TSV containing the following columns.

• strand : on which strand the comparison took place
• td_score : threshold distance score for the signal comparison
• kl_divergence : kullback leibler divergence for the signal comparison
• bayesian_p : p-value for the signal comparison
• signal_type : classification into "mod" for modification and "mut" for mutation
• ref_1 : contig name of sample 1
• pos_1 : position in contig of sample 1 (0-based)
• base_1 : base at the position of sample 1
• motif_1 : motif around the base at the position of sample 1
• signalmean1 : mean of the signal distribution at the position of sample 1
• signalstd1 : standard deviation of the signal distribution at the position of sample 1
• ndatapoints1 : number of data points that formed the signal distribution
• containeddatapoints1 : number of data points withtin 3 standard deviations around the mean
• nsegments1 : number of segments from nanopolish eventalign that formed the signal distribution
• containedsegments1 : number of segments within 3 standard deviations around the mean
• nreads1 : number of reads (coverage) that formed the signal distribution

same for second sample: - ref2, pos2, base2, motif2, signalmean2, signalstd2, ndatapoints2, containeddatapoints2, nsegments2, containedsegments2, nreads2

• 11: Concatenating both reference files failed
• 12: Building mafft alignment failed
• 13: Running nanosherlock of the first sample failed
• 14: Running nanosherlock of the second sample failed
• 15: Number of provided reference sequences is not equal 1 or 2
• 16: Unknown pore type
• 17: Error in multiprocessing signal comparison

  - 18: Error in magniplot

Errors of first sample: - 119: Cannot basecall .slow5/.blow5 with guppy - 120: Could not find raw data or unknown file format - 121: Guppy basecalling failed - 122: mapping failed - 123: Samtools indexing failed - 124: f5c index failed - 125: f5c eventalign failed - 126: Could not find provided fastq files

- 127: f5c eventalign file is empty

Errors of second sample - 219: Cannot basecall .slow5/.blow5 with guppy - 220: Could not find raw data or unknown file format - 221: Guppy basecalling failed - 222: mapping failed - 223: Samtools indexing failed - 224: f5c index failed - 225: f5c eventalign failed - 226: Could not find provided fastq files - 227: f5c eventalign file is empty

The -e parameter of nanosherlock specifies the leading number of the error code. Default is 0. - 019: Cannot basecall .slow5/.blow5 with guppy - 020: Could not find raw data or unknown file format - 021: Guppy basecalling failed - 022: mapping failed - 023: Samtools indexing failed - 024: f5c index failed - 025: f5c eventalign failed - 026: Could not find provided fastq files - 027: f5c eventalign file is empty