utils.py 文件源码

def iterative_consensus(sequences, program='muscle-medium', threshold=0.6,
        subsample_size=200, maximum_subsample_size=1600):
    """
    Compute a consensus sequence of the given sequences, but do not use all
    sequences if there are many: First, try to compute the consensus from a
    small subsample. If there are 'N' bases, increase the subsample size and
    repeat until either there are no more 'N' bases, all available sequences
    have been used or maximum_subsample_size is reached.
    """
    while True:
        sample = downsampled(sequences, subsample_size)
        aligned = multialign(OrderedDict(enumerate(sample)), program=program)
        cons = consensus(aligned, threshold=threshold).strip('N')
        if 'N' not in cons:
            # This consensus is good enough
            break
        if len(sequences) <= subsample_size:
            # We have already used all the sequences that are available
            break
        subsample_size *= 2
        if subsample_size > maximum_subsample_size:
            break
    return cons