minSNPs_nanopore

This documents the analysis performed in using minSNPs to analyse Nanopore sequence data. The analysis were performed on a HPC using Slurm scheduler, and the scripts may need to be adjusted accordingly when ran elsewhere.

Testing with minSNPs

Identifying the SNPs for the new PubMLST samples

Steps:

1. Use SNIPPY to call SNPs for the new sample (see Scripts/SNIPPY_RUN.sh).
2. Extract all the SNPs defined in megaalignment from SNIPPY result (see Scripts/extract_megaalignment_snps.R), output of final alignment mega_with_newdata.fasta is in figshare see here.
3. Assign CC metadata to new samples based on most similar sample in megaalignment (see Scripts/assign_cc_meta.R), see Results/new_sample_most_similar_mega.csv for the result containing the most similar sample in megaalignment and the CC to be assigned, Results/disagreement.csv for the samples with CC metadata from PubMLST that is different from the CC to be assigned.
4. Neighbour Joining tree is created with MEGA with mega_with_newdata.fasta as the input and all default parameters, output: Result/mega_with_newdata_NJT.nwk. see here for the interactive tree.

Sampling 400 random SNPs

All the SNPs are scrambled and the first 400 is taken, see Scripts/random_snps_selection.R.

Search string generation

For

• SNPs, see Scripts/generate_snp_search_sequence.R.
• gene sequences, see Scripts/generate_gene_search_sequence.R.

Major lineage assignment

Testing with lab generated Nanopore data

Steps:

1. Scan lab generated Nanopore data Scripts/scan_lab_nanopore.R.
2. Transform most similar isolate to most likely CC and aggregated all results for different number of SNPs or reads used : Scripts/aggregate_lab_result.R.

Testing with simulated long-read data

Steps:

1. Simulate long read with pbsim2 (see Scripts/simulated_long_read_generation.R), see Data/BIGSdb_3343897_1178826571_39602.csv for the list of data download from pubMLST and Results/sim_n_reads.csv output for number of reads generated.
2. Assign most likely CC for tested samples based on SNPs distance: Scripts/assign_cc_meta.R
3. Scan simulated Nanopore data Scripts/scan_simulated_read.R.
4. Transform most similar isolate to most likely CC and aggregated all results for different number of SNPs or reads used : Scripts/aggregate_simulation_result.R.

Gene detection

Testing with lab generated Nanopore data

1. Scan lab generated Nanopore data Scripts/scan_lab_gene.R.
2. Transform most similar isolate to most likely CC and aggregated all results for different number of SNPs or reads used : Scripts/aggregate_lab_gene.R.

Testing with simulated data

1. Scan simulated Nanopore data Scripts/scan_simulated_gene.R.
2. Transform most similar isolate to most likely CC and aggregated all results for different number of SNPs or reads used : Scripts/aggregate_simulation_gene.R.

Comparison

Comparison with Krocus

• Simulated data:
  • Script: Scripts/krocus_comparison.R
  • Result: Results/krocus_full_result.csv and Results/krocus_summary.csv
• Lab generated data:
  • Script: Scripts/lab_krocus_comparison.R
  • Result: Results/lab_krocus_full_result.csv and Results/lab_krocus_summary.csv

Comparison with Sketchy

• Simulated data:
  • Script: Scripts/sketchy_comparison.R
  • Result: Results/sketchy_gene_summary.csv and Results/sketchy_summary.csv
• Lab generated data:
  • Script: Scripts/lab_sketchy_comparison.R
  • Result: Results/lab_sketchy_gene_summary.csv and Results/lab_sketchy_summary.csv