cluster_coarse.py

#!/usr/bin/env python

import operator
import IMP.multifit
from IMP import ArgumentParser

# analyse the ensemble, first we will do the rmsd stuff


def parse_args():
    desc = "A script for clustering an ensemble of solutions"
    p = ArgumentParser(description=desc)
    p.add_argument("-m", "--max", type=int, dest="max", default=999999999,
                   help="maximum number of combinations to consider")
    p.add_argument("assembly_file", help="assembly file name")
    p.add_argument("proteomics_file", help="proteomics file name")
    p.add_argument("mapping_file", help="mapping file name")
    p.add_argument("param_file", help="parameter file name")
    p.add_argument("combinations_file", help="combinations file name")
    p.add_argument("diameter", type=float, help="cluster diameter")
    p.add_argument("cluster_file", help="output clusters file name")
    return p.parse_args()


def run(asmb_fn, proteomics_fn, mapping_fn, align_param_fn,
        comb_fn, diameter, output_comb_fn, max_combs):
    asmb_data = IMP.multifit.read_settings(asmb_fn)
    prot_data = IMP.multifit.read_proteomics_data(proteomics_fn)
    mapping_data = IMP.multifit.read_protein_anchors_mapping(
        prot_data, mapping_fn)
    alignment_params = IMP.multifit.AlignmentParams(align_param_fn)

    # load all proteomics restraints
    align = IMP.multifit.ProteomicsEMAlignmentAtomic(
        mapping_data, asmb_data, alignment_params)
    _ = align.get_model()
    mhs = align.get_molecules()
    ensb = IMP.multifit.Ensemble(asmb_data, mapping_data)
    for i, mh in enumerate(mhs):
        ensb.add_component_and_fits(
            mh,
            IMP.multifit.read_fitting_solutions(
                asmb_data.get_component_header(i).get_transformations_fn()))

    mol_path_centers = []  # save the molecule centers for each path
    # iterate over the molecules
    print("NUMBER OF COMPS:", asmb_data.get_number_of_component_headers())
    for i in range(asmb_data.get_number_of_component_headers()):
        mol_centers = []  # all the centers of a specific molecule
        mh_leaves = IMP.core.get_leaves(mhs[i])
        # iterate over the paths and add the center of the path
        mh_paths = mapping_data.get_paths_for_protein(
            prot_data.get_protein_name(i))
        dummy_comb = []
        for j in range(asmb_data.get_number_of_component_headers()):
            dummy_comb.append(0)
        for j in range(len(mh_paths)):
            dummy_comb[i] = j
            ensb.load_combination(dummy_comb)
            # print IMP.core.XYZs(mh_leaves)
            mol_centers.append(IMP.core.get_centroid(IMP.core.XYZs(mh_leaves)))
            ensb.unload_combination(dummy_comb)
        mol_path_centers.append(mol_centers)
    for i, p in enumerate(mol_path_centers):
        print("number of paths for mol:", i, "is", len(p))
    # load combinations
    combs = IMP.multifit.read_paths(comb_fn)
    comb_centroids = []
    for comb in combs[:max_combs]:
        mh_c = []
        for i in range(len(mhs)):
            mh_c += mol_path_centers[i][comb[i]]
        comb_centroids.append(IMP.algebra.VectorKD(mh_c))
    embed = IMP.statistics.VectorDEmbedding(comb_centroids)
    # TODO - use your RMSD clustering
    bin_cluster = IMP.statistics.create_bin_based_clustering(embed, diameter)
    print("number of clusters:", bin_cluster.get_number_of_clusters())
    cluster_stat = []
    for k in range(bin_cluster.get_number_of_clusters()):
        bc = bin_cluster.get_cluster(k)
        cluster_stat.append([len(bc), k, bc])
    cluster_stat = sorted(
        cluster_stat,
        key=operator.itemgetter(0),
        reverse=True)
    cluster_reps = []
    for ind, [cluster_size, cluster_ind,
              cluster_elems] in enumerate(cluster_stat):
        print("cluster index:", ind, "with", cluster_size, "combinations")
        cluster_reps.append(combs[cluster_elems[0]])
    print("============clustering============")
    print("Number of clusters found " + str(len(cluster_reps)))
    print("==================================")
    IMP.multifit.write_paths(cluster_reps, output_comb_fn)


if __name__ == "__main__":
    args = parse_args()
    run(args.assembly_file, args.proteomics_file, args.mapping_file,
        args.param_file, args.combinations_file, args.diameter,
        args.cluster_file, args.max)