:>"^{@sdZddlmZddlZddlZddlmZddlmZddl m Z ddl m Z dZ d ZGd d d eZGd d d eZddddZdS)zExtract information from alignment objects. In order to try and avoid huge alignment objects with tons of functions, functions which return summary type information about alignments should be put into classes in this module. )print_functionN)Alphabet)IUPAC)Seq) FreqTableg?g?c@s eZdZdZddZdddddd Zddddd d Zd d ZdddZddZ ddZ dddZ ddddZ ddZ ddZddddddddZddddd Zd!d"Zd#d$ZdS)% SummaryInfozCalculate summary info about the alignment. This class should be used to caclculate information summarizing the results of an alignment. This may either be straight consensus info or more complicated things. cCs||_g|_dS)zInitialize with the alignment to calculate information on. ic_vector attribute. A list of ic content for each column number. N) alignment ic_vector)selfrr 8/tmp/pip-build-ww9dw3qa/biopython/Bio/Align/AlignInfo.py__init__'s zSummaryInfo.__init__gffffff?XNrcCsd}|jj}x}t|D]o}i}d} x|jD]} |t| jkr>| j|dkr>| j|dkr>| j||krd|| j|Wg} d} xQ|D]I} || | kr| g} || } q|| | kr| j| qW|rH| dkrH||7}q"t| dkrt| t| |kr|| d7}q"||7}q"W|dkr|j|}t||S)aOutput a fast consensus sequence of the alignment. This doesn't do anything fancy at all. It will just go through the sequence residue by residue and count up the number of each type of residue (ie. A or G or T or C for DNA) in all sequences in the alignment. If the percentage of the most common residue type is greater then the passed threshold, then we will add that residue type, otherwise an ambiguous character will be added. This could be made a lot fancier (ie. to take a substitution matrix into account), but it just meant for a quick and dirty consensus. Arguments: - threshold - The threshold value that is required to add a particular atom. - ambiguous - The ambiguous character to be added when the threshold is not reached. - consensus_alpha - The alphabet to return for the consensus sequence. If this is None, then we will try to guess the alphabet. - require_multiple - If set as 1, this will require that more than 1 sequence be part of an alignment to put it in the consensus (ie. not just 1 sequence and gaps). r-.N) rget_alignment_lengthrangelenseqappendfloat_guess_consensus_alphabetr)r threshold ambiguousconsensus_alpharequire_multiple consensuscon_lenn atom_dict num_atomsrecord max_atomsmax_sizeatomr r r dumb_consensus/s:&     zSummaryInfo.dumb_consensuscCsd}|jj}xWt|D]I}i}d} xl|jD]a} |t| jkr>| j||krd|| j|Wg} d} xQ|D]I} || | kr| g} || } q|| | kr| j| qW|r"| dkr"||7}q"t| dkrat| t| |kra|| d7}q"||7}q"W|dkr|j|}t||S)ahOutput a fast consensus sequence of the alignment, allowing gaps. Same as dumb_consensus(), but allows gap on the output. Things to do: - Let the user define that with only one gap, the result character in consensus is gap. - Let the user select gap character, now it takes the same as input. rrrN) rrrrrrrrr)r rrrrrrr r!r"r#r$r%r&r r r gap_consensusys8     zSummaryInfo.gap_consensuscCs\tj|jj}xD|jD]9}tj|jj}t||jstdqWt |drX|j dk rX||j krXt|t j r|t j j krt j }qXtj }nt|t jr |t jj krt j}qXtj}nKt|t jrO|t jj krCt j}qXtj}n tj}|S)aPick an (ungapped) alphabet for an alignment consesus sequence (PRIVATE). This just looks at the sequences we have, checks their type, and returns as appropriate type which seems to make sense with the sequences we've got. z\Alignment contains a sequence with an incompatible alphabet.lettersN)r_get_base_alphabetr _alphabetrZalphabet isinstance __class__ ValueErrorhasattrr)rZIUPACUnambiguousDNAZ generic_dnaZIUPACUnambiguousRNAZ generic_rnaZ IUPACProteinZExtendedIUPACProteinZgeneric_proteinZsingle_letter_alphabet)r rar#Zaltr r r rs,     z%SummaryInfo._guess_consensus_alphabetc Cs|j|\}}xtt|jD]}xt|dt|jD]d}|j|j|j|j|j|j|jjdd|j|jjdd||}qNWq+W|S)aGenerate a replacement dictionary to plug into a substitution matrix. This should look at an alignment, and be able to generate the number of substitutions of different residues for each other in the aligned object. Will then return a dictionary with this information:: {('A', 'C') : 10, ('C', 'A') : 12, ('G', 'C') : 15 ....} This also treats weighted sequences. The following example shows how we calculate the replacement dictionary. Given the following multiple sequence alignment:: GTATC 0.5 AT--C 0.8 CTGTC 1.0 For the first column we have:: ('A', 'G') : 0.5 * 0.8 = 0.4 ('C', 'G') : 0.5 * 1.0 = 0.5 ('A', 'C') : 0.8 * 1.0 = 0.8 We then continue this for all of the columns in the alignment, summing the information for each substitution in each column, until we end up with the replacement dictionary. Arguments: - skip_chars - A list of characters to skip when creating the dictionary. This defaults to an empty list. For instance, you might have Xs (screened stuff) or Ns, and not want to include the ambiguity characters in the dictionary. rweightg?)_get_base_replacementsrrr_pair_replacementr annotationsget)r Z skip_charsZrep_dict skip_itemsZrec_num1Zrec_num2r r r replacement_dictionarys%#  z"SummaryInfo.replacement_dictionaryc Csxtt|D]}||}y||} Wntk rI|SYnX||kr| |kry||| f||7 s  D