NB: this is deprecated code, but some may find use of it.

This program automates the running of processed in performing haplotype reconstruction on longitudinal deep sequencing samples, by analysing co-varying variants in a probabilistic framework.

• Git clone this directory into folder
• Save n bam files into <project_folder>/01_bam/
• Save a .txt file including a newline separated list of your .bam files, in order of time (if you have it) into 01_bam/
• Save a reference .fasta file in 01_bam/
• Edit pun_of_harold.py instructions:
  • max-haplo - The maximum number of haplotypes to attempt in Haplotype_Reconstruction
  • alpha-frac - The fraction of variable sites to include in error estimates,
    • 0.01 is recommended. 0.1 for small genomes
    • Larger is better, Larger is slower
  • Command - the comm and requires to submit jobs in a cluster i.e. sh, qsub
  • conda_activate_loc - the location of the binary to run anaconda
  • conda_env - location of the conda environment to use in job submission
    • punofharold.yml is compatible
  • java - the command required to run java on your system

• packaged in here is a punofharold.yml file, use this to create a conda environment for runtime &/or job submission

The process will run through from num_haplo = 2-max-haplo, running HAROLD. Formatiing outputs & piping into the num_haplo + 1 run. Between num_haplo = 4 to max-haplo ifthe gradient of Log Likelihood for the most recent 3 runs is at any point negative; HAROLD jobs will cease. A summary table & plot will be produced. For each BAM file in 01_bam/ haplotype refinement will be initialised, running each in parallel.