CyclomicsManuscript README

Data analysis scripts used to process and plot the data for the manuscript of CyclomicsSeq.

Date

1 July 2020

Author

Myrthe Jager (contact: m.jager-2@umcutrecht.nl)
Alessio Marcozzi (contact: alessio.marcozzi@gmail.com)

Description

Here, you can find all code (R scripts and Python notebooks) that can be used to reproduce the results of all the paper Figures and most Supplementary Figures of:
"Accurate detection of circulating tumor DNA using nanopore consensus sequencing".

Data

Processed data DOI: 10.5281/zenodo.3925250

Instructions for use

A) Start up (estimated time < 1 hour)

1. Download and install required packages (see below @ Software dependencies, R packages and Python libraries)
2. Download the processed data from Zenodo (DOI: 10.5281/zenodo.3925250)
3. Create output directory /Results/ in the main downloaded folder 'Cyclomics_manuscript' with subfolders:

• /10reps/
• /10reps/basetype/
• /10reps/extra/
• /10reps/insert/
• /consensus_loop/
• /consensus_loop/backbone/
• /consensus_loop/insert/
• /consensus_loop/all/
• /consensus_loop/all/error/
• /consensus_loop/all/qscore/
• /patients/
• /patients/patienta/
• /patients/patientb/
• /patients/patientc/

4. Add dir for each script used in steps B and C (except functions.R) @ GET STARTED step 2: directory of downloaded processed data (once per script)

B) snFP and deletion analyses (estimated time ~ 1 hour)

5. Run functions.R in RStudio
6. Run consensus_loop.R in RStudio to get & plot (~ 35 minutes runtime):

• Figure 3a
• Figure 5a
• Supplementary figures: 4a, 7a, 7b, 8a, 8b and 9b
• Analyses of single-nucleoide FP rate and deletion rate per number of repeat for 1-39 and 40+ repeats
• Separate analyses for entire runs and per insert, and for error and qscore
• Coverage of all runs per number of repeat for 1-39 and 40+ repeats

7. Clear RStudio workspace
8. Run functions.R in RStudio
9. Run 10reps.R in RStudio to get & plot (~ 25 minutes runtime):

• Figure 3 b,c,d
• Figure 5 b,c
• Supplementary figures: 4b-g, 5b-e, 6, 7c-f and 8c-f
• Analyses of single-nucleotide FP rate and deletion rate per nucleotide position and per run (for 10+ repeats)
• Nucleotide positions which require snFP forward/reverse correction
• Forward/reverse corrected analyses of single-nucleotide FP rate and deletion rate per nucleotide position and per run (for 10+ repeats)
• Percentage of bases with <0.1%, 0.1-1.0 and >1.0% snFP and deletion rates per run

10. Clear RStudio workspace

C) Patients (estimated time ~ 2 minutes)

11. Run functions in RStudio
12. Run patientA.R in RStudio to get & plot (~30 seconds runtime)

• Figure 6 a,d

13. Clear RStudio workspace
14. Run functions
15. Run patientB.R in RStudio to get & plot (~30 seconds runtime)

• Figure 6 b,e

16. Clear RStudio workspace
17. Run functions
18. Run patientC.R in RStudio to get & plot (~30 seconds runtime)

• Figure 6 c,f

19. Clear RStudio workspace

D) CyclomicsSeq run-stats (estimated time < 1 hour)

The notebook stats_from_structure.ipynb was used to generate the plots of Figure 1 and the ones in SuppementaryData.zip.

1. Start a jupyter notebook server and open stats_from_structure.ipynb
2. Set the value of the data_folder variable to match the location where you have downloaded the processed data from Zenodo (DOI: 10.5281/zenodo.3925250)
3. Optionally, modify the samples variable to include/exclude specific runs. The default value include all the runs published in "Accurate detection of circulating tumor DNA using nanopore consensus sequencing"
4. Specify the value of save_plots_folder, which indicates the folder where the plots will be saved
5. Run all the cells

E) TP53 exons coverage (estimated time < 10 minutes)

The notebook TP53_panel_coverage.ipynb was used to generate the plots of Figure 2.

1. Start a jupyter notebook server and open TP53_panel_coverage.ipynb
2. Set the value of the data_folder variable to match the location where you have downloaded the processed data from Zenodo (DOI: 10.5281/zenodo.3925250)
3. Set the value of the output_folder variable which indicates the folder where the plots will be saved
4. Optionally, modify the samples variable to include/exclude specific runs
5. Run all the cells

F) False positive rates on TP53 COSMIC mutations (estimated time < 10 minutes)

The notebook COSMIC_analysis.ipynb was used to generate the plots of Figure 4.

1. Start a jupyter notebook server and open COSMIC_analysis.ipynb
2. Set the value of the data_folder variable to match the location where you have downloaded the processed data from Zenodo (DOI: 10.5281/zenodo.3925250)
3. Optionally, modify the samples variable to include/exclude specific runs
4. Optionally, modify the save_as variable to change the name and the extension of the output file
5. Run all the cells

G) Visualize the structure of a CyclomicsSeq read (estimated time < 1 hour)

The CyclomicsSeq concatemers can be visualized using plot_read_structure.ipynb

1. Start a jupyter notebook server and open plot_read_structure.ipynb
2. Set the value of the data_folder variable to match the location where you have downloaded the processed data from Zenodo (DOI: 10.5281/zenodo.3925250)
3. Set the value of the sample_name variable to specify the run to be analyzed
4. Optionally, set the other input parameters specified in the second cell of the notebook
5. Run all the cells

Software dependencies

R version 3.5.1 RSrudio version 1.1.456

Python version 3.6 jupyterlab version 1.2.6

R packages

• cowplot 0.9.3
• ggplot2 3.0.0
• gridExtra 2.3
• reshape2 1.4.3
• scales 0.5.0

Python libraries

• matplotlib 3.1.3
• numpy 1.18.5
• pandas 1.0
• scipy 1.4
• biopython 1.7