Xf_/hdZddlmZGddejeZGddeZd dZd Zd S) zParsing TRANSFAC files.)motifsc2eZdZdZhdZhdZfdZxZS)MotifaStore the information for one TRANSFAC motif. This class inherits from the Bio.motifs.Motif base class, as well as from a Python dictionary. All motif information found by the parser is stored as attributes of the base class when possible; see the Bio.motifs.Motif base class for a description of these attributes. All other information associated with the motif is stored as (key, value) pairs in the dictionary, where the key is the two-letter fields as found in the TRANSFAC file. References are an exception: These are stored in the .references attribute. These fields are commonly found in TRANSFAC files:: AC: Accession number AS: Accession numbers, secondary BA: Statistical basis BF: Binding factors BS: Factor binding sites underlying the matrix [sequence; SITE accession number; start position for matrix sequence; length of sequence used; number of gaps inserted; strand orientation.] CC: Comments CO: Copyright notice DE: Short factor description DR: External databases [database name: database accession number] DT: Date created/updated HC: Subfamilies HP: Superfamilies ID: Identifier NA: Name of the binding factor OC: Taxonomic classification OS: Species/Taxon OV: Older version PV: Preferred version TY: Type XX: Empty line; these are not stored in the Record. References are stored in an .references attribute, which is a list of dictionaries with the following keys:: RN: Reference number RA: Reference authors RL: Reference data RT: Reference title RX: PubMed ID For more information, see the TRANSFAC documentation. >BFBSCCDRDTHCHPOV>RARLRTRXc t|}n=#t$r0ttj||}YnwxYw|S)N)super __getitem__ TypeErrorrr)selfkeyvalue __class__s 3lib/python3.11/site-packages/Bio/motifs/transfac.pyrzMotif.__getitem__Fsd ?GG'',,EE ? ? ?&,--99#>>EEE ? s!%7AA)__name__ __module__ __qualname____doc__multiple_value_keysreference_keysr __classcell__)rs@rrr s_00dKJJ.--NrceZdZdZdZdZdS)RecordzStore the information in a TRANSFAC matrix table. The record inherits from a list containing the individual motifs. Attributes: - version - The version number, corresponding to the 'VV' field in the TRANSFAC file; cd|_dS)zInitialize the class.N)versionrs r__init__zRecord.__init__Ys  r"c t|S)z'Turn the TRANSFAC matrix into a string.)writer's r__str__zRecord.__str__]sT{{r"N)rrrrr(r+r"rr$r$Ns<r"r$Tc B i}g}d}t}|D]}|}|s|dd}|d}|r&t|dkrt d|dt|dkrJ|d} |r.|ddst d|d|d kr | |_n+|d vr&i}| dd gd krt d |d|d} dD]} g|| <|D]}|}|dd}|d}t|dkrJ|d} |r.|ddst d|d t|} n#t$rYnwxYw| dkr| dkr|rt d|n| dz } | | krt d|d|rUt|dkrt d| dd| dd|ddt|dkrt d|d| dd } t| d krt d|dtd| D]-\} }||  t|.|dkr|dkr| d\}}}|ddkrt d|d |d|d!d"krt d|d#|dt|dd!}t||dz kr,t d|dd$t|dzdd|dd|| i}| ||d%krJ|Btd|&}| |||_ | |i}g}8|tjvr| ||<M|tjvr&||vrg||<|| | | ||<|S)'z4Parse a transfac format handle into a Record object.NrzAThe key value of a TRANSFAC motif line should have 2 characters:"" zKA TRANSFAC motif line should have 2 spaces between key and value columns: "VV)P0PO)ACGTzA TRANSFAC matrix "z(" line should be followed by "A C G T": ACGTzaA TRANSFAC matrix should start with "01" as first row of the matrix, but this matrix uses "00": "zKThe TRANSFAC matrix row number does not match the position in the matrix: "zLA TRANSFAC matrix line should have a 2 digit key at the start of the line ("02dz"), but this matrix uses "dz": "sz".z7A TRANSFAC matrix line should have a key and a value: "zQA TRANSFAC matrix line should have a value for each nucleotide (A, C, G and T): "XXRN;[z The index "z2" in a TRANSFAC RN line should start with a "[": "]z0" in a TRANSFAC RN line should end with a "]": "zP" of the TRANSFAC RN line does not match the current number of seen references "//)alphabetcounts)r$stripsplitlen ValueError partitionr&intzipappendfloatrupdate referencesr r)handlestrict annotationsrQrFrecordline key_valuerrlengthcivaluesvindex separator accession referencemotifs rreadrbbsKJ F XXFx%x%zz||  JJtQ'' l  ""  3xx1}}      y>>Q  aL&&((E ~~d++A.$$ $$$ $;;"FNN L F{{}}RaR $8$8$888 5#55.255F  q 0 /0 /zz|| JJtQ// l((**y>>Q&&%aL..00E#~~d33A6",!LDH!L!L!L##CAA!EEQ;;166(RKORR aKF;;$4,0444 3xx1}}(J?@IJJ67IJJ?CIJJJ 9~~**(0(,000rr*v;;!##$A9=AAA ////DAq1I$$U1XX..../ 4<<  D[[*///#*>*> 'E9iQx3 -%--%)---RyC -%--%)---ad $$E:%!)++ =%N==J!+<==37<=== e I   i ( ( ( ( D[[!vf=== [)))#-  e$$$KJJ E( ( ("IcNN E- - -+%%#% C   # #E * * * *$K   Ms&G66 HHc g} |j}|d|z}||n#t$rYnwxYwtj}d}|D]!g}|D]}d}|D]} | dkr׉j} | dkrj} tj} d dg| z} || t| D]md d gd | Dzd z} | td zgfd | Dz| gzz} || nd}no  | }n#t$rd}YnwxYw|D| |vr"|D]}| d|} || n| d|} || d}| dkr] j }d}|D]=}|D]8} | | }|| d|} || d}9>#t$rYwxYw|rd} || d} || d |dz}||#d |}|S)z7Write the representation of a motif in TRANSFAC format.Nz VV %s XX // ))ACAS)ID)r CO)NA)DE)TY)OSOC)r r )r)r3)BA)r)r)r )r PVFr3rz  z%02.dcg|]}dS)z%6.20gr,).0_s r zwrite..s1L1L1Lq(1L1L1Lr"z %sr.c6g|]}j|Sr,)rF)rqrrrZras rrszwrite..s$CCCau|Aq1CCCr"Tr1rn)r?rrrrr>rD )r&rNAttributeErrorrrrXdegenerate_consensussortedrEjoinrangetuplegetrQ)rblocksr&blockrsectionslinessectionblankrrXsequencelettersrVrr\rQkeysr`textrZras @@rr*r*s F!.      MM%        3H"@@9 #9 #GE4 -4 -$;;"\F{{ $9H$U^44G#==$')9::DLL&&&"6]] + +HHgY1L1LG1L1L1L%LMM() $eUGCCCCC7CCCD'{m,''   T**** EE% % #)%%% $%("555%*33*-}}}} % T 2 2 2 23'*#4#4U#4#4D!LL... $$;; -%*%5  >)3--I'+--(1 c(:(:#(=$,*-'8'8'8'8 % T 2 2 2(, -- *   # T""" T %  4' e 776??D Ks0* 779E E E .G99 H H N)T) rBiorrdictlistr$rbr*r,r"rrs>>>>>FL$>>>BT(Dgggggr"