hiv_polyA_struct_phylo_analysis

Goal: to analyze the structural conservation of the 5' polyA hairpin the HIV genome.

Code used for the analysis in Yasin, S., Lesko, S., Kharytonchyk, S., Brown, J., Chaudry, I. et al. Role of RNA structural plasticity in modulating HIV-1 genome packaging and translation. Proc. Natl. Acad. Sci. 121, (2024). https://doi.org/10.1073/pnas.2407400121

Pipeline

There are three main components of the pipeline and each is seperated into a seperate directory in the repository. Each component is described further.

• polyA_detection
• make_tree-post_detection_analysis
• concensus_structure-post_detection_analysis

---

polyA_detection

Goal: Parse out the putative polyA hairpin from a set of target HIV genomes using de novo discovery method described below.

Requirements: viennaRNA, clustalo

Input file (parameters are explained in '<< >>' - do not include brackets in file when running):

{
    "paths": {
        "full_genomes_path": ""<<Path to a JSON file with the target genomes. See example for input format.>>,
        "rna_fold_path": ""<<Path to viennaRNA executable (RNAfold)>>,
        "clustal_path": "" <<Path to clustalo executable>>
    },
    "initial_seg": {
        "ext_range": 75 <<# of basepairs to extend up and down stream from the AAUAAA signal to pull the initital fragment to pull the polyA from>>
    },
    "minimization": {
        "min_len": 35, <<Minimum length of fragment to be conisdered a potential putative polyA hairpin>>
        "max_len": 50, <<Maximum length of fragment to be conisdered a potential putative polyA hairpin>>
        "max_unbp_ratio": 0.45, <<Ratio representing the max percentage of unbase paired residues>>
        "hwd_limit": 0.25 <<A restraint on how far from the center of the hairpin the AAUAAA can be for a potential putative polyA hairpin.>>
    },
    "output": {
        "folder": "../output/", << >>
        "polyA_output": "polyA.fasta" << >>
    },
    "strains_force_include":["KM390026", "K03455", "AB286862", "U51188", "X04415", "MH705144", "AB485658", "MH705163", "KC156211"] <<Strains that will be forced into the final output - regardless of whether or not they are represenative of their subtype.>>
}

This will output representative strains from the major HIV subtypes.

make_tree-post_detection-analysis

Goal: Using the output from the polyA detection to construct an annotated phylogenetic tree.

Requirements: phyml, clustalo

Input file (parameters are explained in '<< >>' - do not include brackets in file when running):

{   
    "paths":{
        "strain_info":"" <<Path to the output from the polyA detection pipeline.>>
    },
    "alignment_tree": {
        "clustal_path": "", <<Path to the clustalo exectuable.>>
        "phyml_path": "" <<Path to the phyml executable>>
    },
    "params":{
        "bootstrap":1, <<# of bootstrap replicates for the MLH tree>>
        "phylo_gene":"env" <<Target gene to be used to construct the phylogeny>>
    }
}

This will output a SVG with the tree.

concensus_structure-post_detection_analysis

Goal: Using the output from the polyA detection to determine the concensus structure and visualize the variation via frequency plot.

Note: the concensus structure is determined using the locarna package.

Input file (parameters are explained in '<< >>' - do not include brackets in file when running):

{   
    "concensus_sequence":"CACUGCUUAAGCCUCAAUAAAGCUUGCCUUGAGUGCUUHAAGURGUG", <<Concensus sequence outputted by locarna - ignoring gaps>>
    "concensus_vstr":"(((((((((((((((((...........))))).)))).))))))))", << Vstr for the concensus sequence outputted by locarna - ignoring gaps>>
    "target_strains_path":"", <<Path to the JSON file with all the strains to be used in the frequency calculations>>
    "colors":{"AU": "#f5c31d", "GC": "#3e4b94", "GU": "#852a2a"} <<Colors to use for the occurence of each base pair type in the frequecy bars>>
}

This will output a SVG with a bar per position of the hairpin alignments showing the frequncy of each base pair type that is present there.

Note: Hairpins are aligned and compared using a modified Needleman-Wunch algorithm. See paper methods for more details.

---

Usage

Set up a conda environment using the mfs_lab.yml configuration file. The usage for each component of the pipline follows the format [main_file].py input.json