o Rŀg@s"dZddlZddlZddlmZddlmZddlmZGdddeZGdd d Z Gd d d Z Gd d d eZ ddZ ddZ ddZddZddZddZddZddZddZd d!Zd"d#Zd$d%Zd&d'Zd(d)Zd*d+Zd,d-Zd.d/Zed0krdd1lmZedd2dSdS)3aNCode to work with the sprotXX.dat file from SwissProt. https://web.expasy.org/docs/userman.html Classes: - Record Holds SwissProt data. - Reference Holds reference data from a SwissProt record. Functions: - read Read one SwissProt record - parse Read multiple SwissProt records N)Position) SeqFeature)SimpleLocationcs&eZdZdZddfdd ZZS)SwissProtParserErrorz1An error occurred while parsing a SwissProt file.Nlinecstj|||_dS)z=Create a SwissProtParserError object with the offending line.N)super__init__r)selfrargs __class__J/var/www/html/myenv/lib/python3.10/site-packages/Bio/SwissProt/__init__.pyr s  zSwissProtParserError.__init__)__name__ __module__ __qualname____doc__r __classcell__rrr rrsrc@eZdZdZddZdS)Recorda Holds information from a SwissProt record. Attributes: - entry_name Name of this entry, e.g. RL1_ECOLI. - data_class Either 'STANDARD' or 'PRELIMINARY'. - molecule_type Type of molecule, 'PRT', - sequence_length Number of residues. - accessions List of the accession numbers, e.g. ['P00321'] - created A tuple of (date, release). - sequence_update A tuple of (date, release). - annotation_update A tuple of (date, release). - description Free-format description. - gene_name A list of dictionaries with keys 'Name', 'Synonyms', 'OrderedLocusNames' and 'ORFNames'. - organism The source of the sequence. - organelle The origin of the sequence. - organism_classification The taxonomy classification. List of strings. (http://www.ncbi.nlm.nih.gov/Taxonomy/) - taxonomy_id A list of NCBI taxonomy id's. - host_organism A list of names of the hosts of a virus, if any. - host_taxonomy_id A list of NCBI taxonomy id's of the hosts, if any. - references List of Reference objects. - comments List of strings. - cross_references List of tuples (db, id1[, id2][, id3]). See the docs. - keywords List of the keywords. - features List of tuples (key name, from, to, description). from and to can be either integers for the residue numbers, '<', '>', or '?' - protein_existence Numerical value describing the evidence for the existence of the protein. - seqinfo tuple of (length, molecular weight, CRC32 value) - sequence The sequence. Examples -------- >>> from Bio import SwissProt >>> example_filename = "SwissProt/P68308.txt" >>> with open(example_filename) as handle: ... records = SwissProt.parse(handle) ... for record in records: ... print(record.entry_name) ... print(record.accessions) ... print(record.keywords) ... print(record.organism) ... print(record.sequence[:20] + "...") ... NU3M_BALPH ['P68308', 'P24973'] ['Electron transport', 'Membrane', 'Mitochondrion', 'Mitochondrion inner membrane', 'NAD', 'Respiratory chain', 'Translocase', 'Transmembrane', 'Transmembrane helix', 'Transport', 'Ubiquinone'] Balaenoptera physalus (Fin whale) (Balaena physalus). MNLLLTLLTNTTLALLLVFI... cCsd|_d|_d|_d|_g|_d|_d|_d|_g|_g|_ g|_ d|_ g|_ g|_ g|_g|_g|_g|_g|_g|_g|_d|_d|_d|_dS)Initialize the class.N) entry_name data_class molecule_typesequence_length accessionscreatedsequence_updateannotation_update description gene_nameorganism organelleorganism_classification taxonomy_id host_organismhost_taxonomy_id referencescommentscross_referenceskeywordsfeaturesprotein_existenceseqinfosequencer rrrr Zs0 zRecord.__init__Nrrrrr rrrrr$s 5rc@r) ReferenceaHolds information from one reference in a SwissProt entry. Attributes: - number Number of reference in an entry. - evidence Evidence code. List of strings. - positions Describes extent of work. List of strings. - comments Comments. List of (token, text). - references References. List of (dbname, identifier). - authors The authors of the work. - title Title of the work. - location A citation for the work. cCs.d|_g|_g|_g|_g|_g|_g|_dS)rN)number positionsr*r)authorstitlelocationr1rrrr s zReference.__init__Nr2rrrrr3ys r3c@seZdZdZdS) FeatureTablea!Stores feature annotations for specific regions of the sequence. This is a subclass of SeqFeature, defined in Bio.SeqFeature, where the attributes are used as follows: - ``location``: location of the feature on the canonical or isoform sequence; the location is stored as an instance of SimpleLocation, defined in Bio.SeqFeature, with the ref attribute set to the isoform ID referring to the canonical or isoform sequence on which the feature is defined - ``id``: unique and stable identifier (FTId), only provided for features belonging to the types CARBOHYD, CHAIN, PEPTIDE, PROPEP, VARIANT, or VAR_SEQ - ``type``: indicates the type of feature, as defined by the UniProt Knowledgebase documentation: - ACT_SITE: amino acid(s) involved in the activity of an enzyme - BINDING: binding site for any chemical group - CARBOHYD: glycosylation site; an FTId identifier to the GlyConnect database is provided if annotated there - CA_BIND: calcium-binding region - CHAIN: polypeptide chain in the mature protein - COILED: coiled-coil region - COMPBIAS: compositionally biased region - CONFLICT: different sources report differing sequences - CROSSLNK: posttransationally formed amino acid bond - DISULFID: disulfide bond - DNA_BIND: DNA-binding region - DOMAIN: domain, defined as a specific combination of secondary structures organized into a characteristic three-dimensional structure or fold - INIT_MET: initiator methionine - INTRAMEM: region located in a membrane without crossing it - HELIX: alpha-, 3(10)-, or pi-helix secondary structure - LIPID: covalent binding of a lipid moiety - METAL: binding site for a metal ion - MOD_RES: posttranslational modification (PTM) of a residue, annotated by the controlled vocabulary defined by the ptmlist.txt document on the UniProt website - MOTIF: short sequence motif of biological interest - MUTAGEN: site experimentally altered by mutagenesis - NON_CONS: non-consecutive residues - NON_STD: non-standard amino acid - NON_TER: the residue at an extremity of the sequence is not the terminal residue - NP_BIND: nucleotide phosphate-binding region - PEPTIDE: released active mature polypeptide - PROPEP: any processed propeptide - REGION: region of interest in the sequence - REPEAT: internal sequence repetition - SIGNAL: signal sequence (prepeptide) - SITE: amino-acid site of interest not represented by another feature key - STRAND: beta-strand secondary structure; either a hydrogen-bonded extended beta strand or a residue in an isolated beta-bridge - TOPO_DOM: topological domain - TRANSIT: transit peptide (mitochondrion, chloroplast, thylakoid, cyanelle, peroxisome, etc.) - TRANSMEM: transmembrane region - TURN: H-bonded turn (3-, 4-, or 5-turn) - UNSURE: uncertainties in the sequence - VARIANT: sequence variant; an FTId is provided for protein sequence variants of Hominidae (great apes and humans) - VAR_SEQ: sequence variant produced by alternative splicing, alternative promoter usage, alternative initiation, or ribosomal frameshifting - ZN_FING: zinc finger region - ``qualifiers``: a dictionary of additional information, which may include the feature evidence and free-text notes. While SwissProt includes the feature identifier code (FTId) as a qualifier, it is stored as the attribute ID of the FeatureTable object. N)rrrrrrrrr9sr9ccsPt|}z t|}|sW||ur|dSdS|Vq||ur'|ww)zRead multiple SwissProt records from file. Argument source is a file-like object or a path to a file. Returns a generator object which yields Bio.SwissProt.Record() objects. TN)_open_readclosesourcehandlerecordrrrparses  rAcCspt|}z)t|}|stdz t|Wtdty-|YW||ur,|SSw||ur7|ww)zRead one SwissProt record from file. Argument source is a file-like object or a path to a file. Returns a Record() object. zNo SwissProt record foundz$More than one SwissProt record found)r:r; ValueErrornext StopIterationr<r=rrrreads"    rEcCsJzt|}|WSty$|}|ddkr|YStj|ddYSw)NrrASCII)encoding)open TypeErrorrEio TextIOWrapper)r>r?rrrr:s r:c Cs`d}d}zt|}Wn ty|YSw|dd|dd}}|dkr/td|dt}t||g}|D]j}|dd|dd}}|rW|d|}d}|dkrj|d d }|j|q;|d krtt ||q;|d kr|j | q;|d kr|dkr|j dq;t|j dkr|j d|j d|d7<q;|dkr|j |q;|dkr|j|dd7_q;|dkr|dd }|j|q;|dkrt||q;|dkrt||q;|dkrt} t| ||j | q;|dkr|js Jd|jdj |q;|dkr-|js"Jd|jd} t| |}q;|dkrE|js:Jd|jd} t| |q;|dkr^|jsRJd|jd} | j |q;|d krw|jskJd!|jd} | j |q;|d"kr|jsJd#|jd} | j |q;|d$kr|jsJd%|jd} | j |q;|d&krt||q;|d'krt||q;|d(krt||q;|d)krt ||q;|d*krt!||q;|d+kr|}t|d,ksJd-|t"|d.t"|d/|df|_#q;|d0kr| |$ddq;|d1krt%|d&|j |_ d&|j|_|j|_|jD]S} d&| jd | _| jr| jd} | jd.dD]} | 'd2sa| d7} | | 7} qU| d } | (d3r~| 'd3r~| d.d} nd} | | _d&| j| _q9d&||_)|S|d4krq;td5|d6|d|rt*d7dS)8NrIDzFailed to find ID in first liner AC;; DTDEGNandrOSOGOC;.OXOHRNRPzRP: missing RNRCzRC: missing RNRXzRX: missing RNRLzRL: missing RNRAzRA: missing RNRGzRG: missing RNRTzRT: missing RNCCDRPEKWFTSQzI don't understand SQ line z z//-"z**zUnknown keyword z foundzUnexpected end of stream.)+rCrDrstriprr_read_idsplitrextend_read_dtr!appendstripr"lenr#r$r%_read_ox_read_ohr3_read_rnr)r5_read_rc_read_rxr8r6r7_read_cc_read_dr_read_pe_read_kw_read_ftintr/replace_read_gnjoinendswith startswithr0rB) r?r@unreadrkeyvalue_sequence_linesrcols referencer7fragmentrrrr;s                                     "           r;cCst|jD]8\}}|dd}i}|D]"}|dd\}}|dkr*||d<q|dvs0J|d||<q||j|<qdS)NrR=rmName)SynonymsOrderedLocusNamesORFNames, ) enumerater"rqrsrw)r@itexttokensr"tokenrrrrrrs   rcCs|dd}t|dkr+|d|_|dd|_|dd|_t|d|_n$t|dkrI|d|_|dd|_d|_t|d|_ntd|d d }|j|vrbd |j}t||d |jd vrsd |j}t||d dS)NrMrrmrQrLrnzID line has unrecognised formatr)STANDARD PRELIMINARYIPIReviewed UnreviewedzUnrecognized data class )NPRTzUnrecognized molecule type ) rsrxrrqrrrrr)r@rrallowedmessagerrrrrs(           rrc Cs|dd}|}|}d|vsd|vsd|vr|}d}tt|D] }d||vr2|}q(|dks>Jd||d }||d } | d krPd} n d | vrW| } nt| } |d} d|vrj| | f|_dSd|vru| | f|_dSd|vr| | f|_dSt d |dd|vsd|vsd|vrzd} |dd D] } | rt| } qWn t yd} Ynw|dd } d|vr| | f|_dSd|vr| | f|_dSd|vr| | f|_dSt d |dt d|d)NrMCREATEDzLAST SEQUENCE UPDATEzLAST ANNOTATION UPDATErWzREL.rz Could not find Rel. in DT line: rm,r.zUnrecognised DT (DaTe) linerzINTEGRATED INTOzSEQUENCE VERSIONz ENTRY VERSION INTEGRATEDzFailed to parse DT (DaTe) line) upperrqrsrangerxrrrr risdigitrB) r@rruprlineruprcols rel_indexindex version_index str_versionversiondatesrrrrusd       rucCsx|dd}|jr|ddd}n|dddd\}}|dks1Jd||j|ddS) N{rrMrQr NCBI_TaxIDzUnexpected taxonomy type r)rsr&rqrt)r@ridsdescrrrrry3s  rycCs|dddsJd||dd}|ddkr%|ddks)J||ddd\}}|j||j|dS) NrMz NCBI_TaxID=z Unexpected rWrrQrm)rrqcountrsr(rvrwr')r@rtaxidnamerrrrzKs "rzcCs|dd}|d}|dr|dsJd|t|dd|_t|dkrK|d}|dr8|ds?Jd ||ddd |_dSdS) Nrmr[]zMissing brackets rWr}zMissing braces |)rsrrrr4rxevidence)rrnwordsr4rrrrr{Us " r{c Cs|d}|ddkrd}n |dd|d}}|D]=}|s"dS|d}|dkrG|d|||dd}}||f}|j|q|jd}|d|}||jd<q|S)NrQrWrrrrmrO)rsfindlstripr*rv) rrrrcolrrrcommentrrrr|fs      r|cCs|dd}d}d|vrT|d}dd|D}dd|D}|D]1}|d}t|d ks2|d kr6d }nt|d ksCJd ||j|d |ddfq!n |d}t|d krbd }n|j|d d|ddf|rd dl}d dlm}| d||dSdS)Nz [NCBI, ExPASy, Israel, Japan]rFrrRcSsg|]}|qSr)rw.0xrrr z_read_rx..cSsg|]}|r|qSrrrrrrrrrL)DOIrTzI don't understand RX line rrmrQr)BiopythonParserWarningzPossibly corrupt RX line ) rrsrxr)rvrqwarningsBiorwarn)rrrrrrrrrrrr}|s.       $ r}cCsr|dd|dd}}|dkr|j|dS|dkr7|js*|j|dS|jdd|7<dSdS)NrM z-!-z rWrO)rqr*rv)r@rrrrrrr~sr~cCs$|dd}|jt|dS)NrrR)rqrsr+rvtuple)r@rrrrrrsrcCs|d}t|d|_dS)N:r)rsrr.)r@rperrrrs rcCsD|ddD]}|dr|ddd}|j|qdS)Nr[rRrrrmr)rqrsrrsplitr,rvrw)r@rvalrrrrs  rcCs|dd}|r|dddkrI|dd}z |d\}}Wn ty.d}Ynwz |d\}}Wn tyE|}d}Ynwi}n|d d }|dd }d}|d d }d|i}t|d}|dkrv|d}nt|}t|||d}t||d|d} |j | dS|jd} |dd dkr|d d }| dr|ddd| _ dS| j d} | dr| |}n| d|}| jdvrz |d\} } Wn tyYn-wd} | d}|dkr| |d} | d|} | dd} | dd} | d| | }|| j d<dS|dd}td|}|r|d}|t|dd}| dsHtd|d kr`| dsWtd!|dd| _ dS| drm|dd}n|dd}|| j vrtd"|d#|| j |<dSt| j }|d}|d}| j |} |d$ks| dr| |}n| d|}| jd%krz |d\} } Wn tyYn-wd} | d}|dkr| |d} | d|} | dd} | dd} | d| | }|| j |<dS)&NrM z Prz..r"Kr!rWrm)ref)r8typeid qualifiersz z/FTId=rrorO)VARSPLICVAR_SEQz -> z (z ^/([a-z_]+)=rrpz!Missing starting quote in featurerzMissing closing quote for idzFeature qualifier z already exists for featurerr)rqrsrBrr fromstringrr9r-rvrrrrrrrrematchgrouprxlistkeys)r@rrr8 isoform_idfrom_resto_resrr!featureold_description first_seq second_seq extra_infoextra_info_posrrqualifier_typerrrrrrs                                     r__main__) run_doctest)verbose) rrJrBio.SeqFeaturerrrrBrrr3r9rArEr:r;rrrruryrzr{r|r}r~rrrrr Bio._utilsrrrrrs@    UM  d . |