better Cyrillic: punctuation, splitting tokens like сізден.Алдын, Num…

…ero sign etc.

current

@@ -1 +1 @@
-v1.3.8
+v1.3.9

v1.3.7/README.txt → v1.3.9/README.txt

@@ -1,5 +1,5 @@
-tok-eng version 1.3.7
-Release date: November 30, 2020
+tok-eng version 1.3.9
+Release date: December 5, 2020
 Author: Ulf Hermjakob, USC Information Sciences Institute
 
 English tokenizer tokenize-english.pl
@@ -43,6 +43,13 @@ one, but the tokenizer leaves 'spectrin-F-actin' unsplit.)
 
 -----------------------------------------------------------------
 
+Changes in version 1.3.9:
+- More improvements in handling of Cyrillic text, 
+  especially punctuation at start end end of words
+  and splitting Cyrillic words such as сізден.Алдын (using capitalization).
+- Better handling of Numero sign, Middle dot, Bullet
+Changes in version 1.3.8:
+- Better handling of Cyrillic text, especially name initials.
 Changes in version 1.3.7:
 - Better handling of Cyrillic text, especially hyphenated tokens.
 - Better handling of some em/en-dashes, replacement character at beginning or end of token.

v1.3.7/lib/NLP/English.pm → v1.3.9/lib/NLP/English.pm

@@ -9,7 +9,7 @@ package NLP::English;
 use File::Basename;
 use File::Spec;
 
-# tok v1.3.7 (November 30, 2020)
+# tok v1.3.9 (December 5, 2020)
 
 $chinesePM = NLP::Chinese;
 $ParseEntry = NLP::ParseEntry;
@@ -1769,6 +1769,9 @@ sub guard_abbreviations_with_dontbreak {
 
 $alpha = "(?:[a-z]|\xCE[\xB1-\xBF]|\xC3[\x80-\x96\x98-\xB6\xB8-\xBF]|[\xC4-\xC9\xCE-\xD3][\x80-\xBF]|[\xD0-\xD3][\x80-\xBF]|\xE0[\xA4-\xA5][\x80-\xBF]|\xE0[\xB6-\xB7][\x80-\xBF])";
 $alphanum = "(?:[a-z0-9]|\xCE[\xB1-\xBF]|\xC3[\x80-\x96\x98-\xB6\xB8-\xBF]|[\xC4-\xC9\xCE-\xD3][\x80-\xBF]|[\xD0-\xD3][\x80-\xBF]|\xE0[\xA4-\xA5][\x80-\xBF]|\xE0[\xB6-\xB7][\x80-\xBF])(?:[-_a-z0-9]|\xCE[\xB1-\xBF]|\xC3[\x80-\x96\x98-\xB6\xB8-\xBF]|[\xC4-\xC9\xCE-\xD3][\x80-\xBF]|\xE0[\xA4-\xA5][\x80-\xBF]|\xE0[\xB6-\xB7][\x80-\xBF])*(?:[a-z0-9]|\xCE[\xB1-\xBF]|\xC3[\x80-\x96\x98-\xB6\xB8-\xBF]|[\xC4-\xC9\xCE-\xD3][\x80-\xBF]|\xE0[\xA4-\xA5][\x80-\xBF]|\xE0[\xB6-\xB7][\x80-\xBF])|(?:[a-z0-9]|\xCE[\xB1-\xBF]|\xC3[\x80-\x96\x98-\xB6\xB8-\xBF]|[\xC4-\xC9\xCE-\xD3][\x80-\xBF]|\xE0[\xA4-\xA5][\x80-\xBF]|\xE0[\xB6-\xB7][\x80-\xBF])";
+$cyrillic = "(?:[\xD0-\xD3][\x80-\xBF])";
+$uc_cyrillic = "(?:\xD0[\x80-\xAF]|\xD1[\xA0\xA2\xA4\xA6\xA8\xAA\xAC\xAE\xB0\xB2\xB4\xB6\xB8\xBA\xBC\xBE]|\xD2[\x80\x8A\x8C\x8E\x90\x92\x94\x96\x98\x9A\x9C\x9E\xA0\xA2\xA4\xA6\xA8\xAA\xAC\xAE\xB0\xB2\xB4\xB6\xB8\xBA\xBC\xBE]|\xD3[\x80\x81\x83\x85\x87\x89\x8B\x8D\x90\x92\x94\x96\x98\x9A\x9C\x9E\xA0\xA2\xA4\xA6\xA8\xAA\xAC\xAE\xB0\xB2\xB4\xB6\xB8])";
+$lc_cyrillic = "(?:\xD0[\xB0-\xBF]|\xD1[\x80-\x9F\xA1\xA3\xA5\xA7\xA9\xAB\xAD\xAF\xB1\xB2\xB5\xB7\xB9\xBB\xBD\xBF]|\xD2[\x81\x8B\x8D\x8F\x91\x93\x95\x97\x99\x9B\x9D\x9F\xA1\xA3\xA5\xA7\xA9\xAB\xAD\xAF\xB1\xB3\xB5\xB7\xB9\xBB\xBD\xBF]|\xD3[\x82\x84\x86\x88\x8A\x8C\x8E\x91\x93\x95\x97\x99\x9B\x9D\x9F\xA1\xA3\xA5\xA7\xA9\xAB\xAD\xAF\xB1\xB3\xB5\xB7\xB9])";
 
 sub normalize_punctuation {
    local($caller, $s) = @_;
@@ -1908,9 +1911,26 @@ sub tokenize {
          $s =~ s/([\xD8-\xDB][\x80-\xBF])([,;:!?.\(\)\[\]\/]|\xD8\x8C|\xD8\x9B|\xD8\x9F|\xD9\xAA|\xC2\xAB|\xC2\xBB|\xE2[\x80-\x9F][\x80-\xBF])/$1 $2/gi; # punctuation includes Arabic ,;?%
          $s =~ s/([,;:!?.\(\)\[\]\/]|\xD8\x8C|\xD8\x9B|\xD8\x9F|\xD9\xAA|\xC2\xAB|\xC2\xBB|\xE2[\x80-\x9F][\x80-\xBF])([\xD8-\xDB][\x80-\xBF])/$1 $2/gi;
       }
-   }
-   $s =~ s/(\d|[a-zA-Z]|[\xD0-\xD3\xD8-\xDB][\x80-\xBF])([-]|\xE2\x80[\x93\x94])([\xD0-\xD3\xD8-\xDB][\x80-\xBF])/$1 \@$2\@ $3/g;
-   $s =~ s/(\d|[a-zA-Z])([\xD8-\xDB][\x80-\xBF])/$1 \@\@ $2/g;
+      $s =~ s/(\d|[a-zA-Z]|[\xD0-\xD3\xD8-\xDB][\x80-\xBF])([-]|\xE2\x80[\x93\x94])([\xD0-\xD3\xD8-\xDB][\x80-\xBF])/$1 \@$2\@ $3/g;
+      $s =~ s/(\d|[a-zA-Z])([\xD8-\xDB][\x80-\xBF])/$1 \@\@ $2/g;
+      $s =~ s/ ($uc_cyrillic\.)($uc_cyrillic\.)($uc_cyrillic\.)($uc_cyrillic$cyrillic{2})/ $1 $2 $3 $4/g; # added Dec. 1, 2020
+      $s =~ s/ ($uc_cyrillic\.)($uc_cyrillic\.)($uc_cyrillic$cyrillic{2})/ $1 $2 $3/g; # added Dec. 1, 2020
+      $s =~ s/ ($uc_cyrillic\.)($uc_cyrillic$cyrillic{2})/ $1 $2/g; # added Dec. 1, 2020
+      $s =~ s/ ((?:[1-9]\d*|I|II|III|IV|V|VI|VII|VIII|IX|X)\.)($uc_cyrillic$cyrillic)/ $1 $2/g; # added Dec. 1, 2020
+      $s =~ s/([,;:()'*%|]|\xE2\x80\x99)($cyrillic)/$1 $2/g; # added Dec. 5, 2020 RIGHT SINGLE QUOTATION MARK (U+2019)
+      $s =~ s/([,;:()'*%|])([-]|\xE2\x80[\x93\x94])($cyrillic)/$1 \@$2\@ $3/g; # added Dec. 5, 2020
+      $s =~ s/([ 0-9.,;:?!()'*|][.+\\]+)($cyrillic)/$1 $2/g; # added Dec. 5, 2020
+      $s =~ s/([ 0-9.,;:?!()'*|][.+\\%]+)([-]|\xE2\x80[\x93\x94])($cyrillic)/$1 \@$2\@ $3/g; # added Dec. 5, 2020
+      $s =~ s/($cyrillic)([,;:()%'*|\\])/$1 $2/g; # added Dec. 5, 2020
+      $s =~ s/($cyrillic)(\.)([-,\/\\])/$1$2 $3/g; # added Dec. 5, 2020
+      $s =~ s/( _+)($cyrillic)/$1 $2/g; # added Dec. 5, 2020
+      $s =~ s/($cyrillic)(_+ )/$1 $2/g; # added Dec. 5, 2020
+      $s =~ s/($cyrillic$lc_cyrillic)(\.)($uc_cyrillic$lc_cyrillic)/$1 $2 $3/g; # added Dec. 5, 2020
+   }
+   $s =~ s/($alpha)(\xC2\xB7|\xE2\x80\xA2)/$1 $2/g; # Middle dot (U+00B7); Bullet (U+2022) added Dec. 5, 2020
+   $s =~ s/(\xC2\xB7|\xE2\x80\xA2)($alpha)/$1 $2/g; # added Dec. 5, 2020
+   $s =~ s/(\xE2\x84\x96)/ $1 /g; # Numero sign (U+2116) added Dec. 5, 2020
+   $s =~ s/(\xE2\x80\xB0)/ $1 /g; # Per Mille sign (U+2030) added Dec. 5, 2020
    $s =~ s/($alpha)(-|\xE2\x80[\x93\x94]|\xEF\xBF\xBD\.*) /$1 \@$2 /g; # added Nov. 30, 2020 # \xEF\xBF\xBD: repl.char.
    $s =~ s/ (-|\xE2\x80[\x93\x94]|\xE2\x99\xA6|\xEF\xBF\xBD)($alpha)/ $1\@ $2/g; # added Nov. 30, 2020
    print "Point K: $s\n" if $local_verbose;

v1.3.7/test/amr-general-corpus.tok-ref → v1.3.9/test/amr-general-corpus.tok-ref

@@ -18622,7 +18622,7 @@ nope ...
 he 's the real deal ! ! !
 Bush asked former <a href="http://en.wikipedia.org/wiki/United_States_Secretary_of_Defense"> Secretary of Defense </a> <a href="http://en.wikipedia.org/wiki/Dick_Cheney"> Dick Cheney </a> to head up a team to help select a running mate for him , but ultimately , Bush decided that Cheney should be the vice presidential nominee .
 While the U.S. Constitution does not specifically disallow a president and a vice president from the same state , it ' does ' prohibit electors from casting both of his or her votes for persons from his or her own state .
-Accordingly , Cheney—who had been a resident of <a href="http://en.wikipedia.org/wiki/Texas"> Texas </a> for nearly 10 years—changed his voting registration back to Wyoming .
+Accordingly , Cheney @—@ who had been a resident of <a href="http://en.wikipedia.org/wiki/Texas"> Texas </a> for nearly 10 years @—@ changed his voting registration back to Wyoming .
 Had Cheney not done this , either he or Bush would have forfeited their electoral votes from the Texas electors .
 i know its wiki , but that is a better source
 <a href="http://en.wikipedia.org/wiki/2000_presidential_campaign"> United States presidential election , 2000 - Wikipedia , the free encyclopedia </a>
@@ -19427,7 +19427,7 @@ Criminals sometimes obtain Medicare numbers for fraudulent billing by conducting
 Excessive or Inappropriate Testing
 Many standard tests can be useful in some situations but not in others .
 The key question in judging whether a diagnostic test is necessary is whether the results will influence the management of the patient .
-Billing for inappropriate tests—both standard and nonstandard—appears to be much more common among chiropractors and joint chiropractic @/@ medical practices than among other health @-@ care providers .
+Billing for inappropriate tests @—@ both standard and nonstandard @—@ appears to be much more common among chiropractors and joint chiropractic @/@ medical practices than among other health @-@ care providers .
 The commonly abused tests include :
 continued at : <a href="http://www.quackwatch.org/02ConsumerProtection/insfraud.html"> Insurance Fraud and Abuse : A Very Serious Problem </a>
 Can you believe there are dumb asses in here so partisan , they are actually defending the insurance companies .
@@ -25870,7 +25870,7 @@ The logging operations removed natural stands of bald cypress and Atlantic white
 A drier swamp and the suppression of wildfires , which once cleared the land for seed germination , created ecological conditions that were less favorable to the survival of cypress stands .
 As a result , plant and animal variety decreased .
 In the late 1990 s , Prince Corporation heir Erik Prince spent part of his inherited wealth to purchase about 6,000 acres ( 24 km2 ) of the GDS to create a " state @-@ of @-@ the @-@ art private training ground for shooters and security operators " .
-[ 2 ] His private military contracting company—Blackwater Worldwide—is named for the peat @-@ colored water of the swamp .
+[ 2 ] His private military contracting company @—@ Blackwater Worldwide @—@ is named for the peat @-@ colored water of the swamp .
 Also nearby are SEAL and Delta Force bases and the Blackwater facilities are rented out to various government agencies for SWAT team training .
 The Great Dismal Swamp National Wildlife Refuge is located not only between two states , but also between two eco @-@ regions , allowing for a wide range of plant and animal species .
 Baldcypress , tupelo , maple , Atlantic white cypress , and pine are the predominant tree species found on the refuge and support the wildlife within .

v1.3.7/test/bio-amr-snt.tok-ref → v1.3.9/test/bio-amr-snt.tok-ref

@@ -3636,7 +3636,7 @@ In the present study , we discovered that Snail is expressed briefly at an early
 Thereafter , Snail disappears and remains absent during subsequent follicle down @-@ growth and maturation .
 This exquisite pattern appears to be functionally relevant since altering it in vivo correspondingly affects features associated with hair bud formation , including down-regulation of E-cadherin , increased proliferation , and repressed terminal differentiation .
 Although the temporal spike of Snail in the hair bud is reflected at the mRNA level and seems to follow Wnt signaling and BMP inhibition , LEF-1 @/@ β-catenin activation does not appear to induce <i> Snail </i> gene expression in embryonic skin keratinocytes .
-In contrast , we provide in vitro , transgenic ( Tg ) , and gene targeting evidence to show that TGF-β2 and small phenotype– and mothers against decapentaplegic @–@ related protein 2 ( SMAD2 ) signaling are upstream inducers of <i> Snail </i> gene expression in skin epithelium .
+In contrast , we provide in vitro , transgenic ( Tg ) , and gene targeting evidence to show that TGF-β2 and small phenotype @– and mothers against decapentaplegic @–@ related protein 2 ( SMAD2 ) signaling are upstream inducers of <i> Snail </i> gene expression in skin epithelium .
 In the absence of TGF-β2 signaling and <i> Snail </i> gene expression , hair placodes can form , but further follicle down @-@ growth is blocked .
 Our studies point to the view that Snail likely functions downstream of cell fate specification , at a stage where the bud begins to exhibit enhanced proliferation and migration .
 Results
@@ -6449,7 +6449,7 @@ However , K-Ras , but not N-Ras or H-Ras , confers stem @-@ like properties on c
 K-Ras-4B , the most highly expressed splice variant of K-Ras , binds calmodulin ; H-Ras and N @- Ras do not ( <ce:cross-ref id="crosref0085" refid="bib74"> Villalonga et al. , 2001 </ce:cross-ref> ) .
 We believe that this unique property of K-Ras-4B confers stem @-@ like properties to cells expressing oncogenic K-Ras-4B proteins ( M. Wang and F.M. , unpublished data ) .
 Analysis of human syndromes caused by germline mutations in H-Ras or K-Ras supports the idea that K-Ras is a stronger oncogene .
-Unexpectedly , humans can tolerate germline @-@ activating mutations in H-Ras—the same activating mutations that drive somatic mutations .
+Unexpectedly , humans can tolerate germline @-@ activating mutations in H-Ras @—@ the same activating mutations that drive somatic mutations .
 Costello syndrome , which is characterized by germline H-Ras mutations , is associated with a broad spectrum of developmental abnormalities and a high risk for rhabdomyosarcomas and neuroblastomas ( reviewed in <ce:cross-ref id="crosref0090" refid="bib55"> Rauen , 2013 </ce:cross-ref> ) .
 It is puzzling that these individuals do not succumb to malignancies associated with sporadic H-Ras mutations ( <ce:cross-ref id="crosref0095" refid="tbl1"> Table 1 </ce:cross-ref> ) .
 Although fully activating alleles of H-Ras can be tolerated , fully activated alleles of K-Ras may not .
@@ -6572,7 +6572,7 @@ Whether the same conclusion can be applied to other mutations such as G12D or G1
 The approach of restoring GTP hydrolysis to mutant proteins received a brief infusion of hope when Scheffzek and colleagues showed that G12V H-Ras could indeed hydrolyze a GTP analog diaminobenzophenone @-@ phosphoroamidate @-@ GTP in which the aromatic amino group mimics the catalytic effects of GAP ’s arginine finger ( <ce:cross-ref id="crosref0280" refid="bib2"> Ahmadian et al. , 1999 </ce:cross-ref> ) .
 A small molecule that provided this local charge might therefore trick mutant Ras into GTP hydrolysis .
 At first sight , the GTD- @/@ GTP @-@ binding site of Ras does not offer any room for such a molecule to bind .
-However , these issues deserve rethinking—perhaps G12D offers more possibilities for this kind of attack than G12V , for example .
+However , these issues deserve rethinking @—@ perhaps G12D offers more possibilities for this kind of attack than G12V , for example .
 Targeting Ras Posttranslational Modification Pathways
 Ras proteins are processed in several steps ( reviewed in <ce:cross-ref id="crosref0285" refid="bib24"> Gysin et al. , 2011 </ce:cross-ref> ) , including farnesylation , proteolytic cleavage at the C terminus by RCE1 , and carboxymethylation by isoprenylcysteine carboxyl methyltransferase ( ICMT ) .
 K-Ras-4A , H-Ras , and N-Ras are further processed by palmitoylation ( <ce:cross-ref id="crosref0290" refid="fig3"> Figure 3 </ce:cross-ref> ) .
@@ -6893,12 +6893,12 @@ Analyses of the relative levels of phosphorylated Akt and ERK1 @/@ ERK2 demonstr
 We noted however , that sorafenib inhibited Akt and ERK1 @/@ ERK2 slightly more effectively than SPP86 under these conditions ( Figure <xref ref-type="fig" rid="Fig4"> 4 </xref>@ E ) .
 These differential effects on PI3K @/@ Akt and MAPK signaling may result may stem from the fact that sorafenib and SPP86 target different kinases at low concentrations .
 The enhanced inhibition of MAPK signaling observed with sorafenib may also result from the fact that it targets both RET and RAF family kinases [ @<xref ref-type="bibr" rid="CR37"> 37 </xref> , <xref ref-type="bibr" rid="CR45"> 45 </xref>@ ] .
-Since these observations suggested that SPP86 disrupts ERα- RET crosstalk , we investigated the effect of SPP86 on the proliferation of MCF7 cells .
-Estrogen deprived and serum starved cells were cultured in the presence of 1 ng @/@ ml β- estradiol ( E2 ) or 10 ng @/@ ml GDNF alone and in combination in the presence of 1 μM SPP86 for 7 days .
+Since these observations suggested that SPP86 disrupts ERα @- RET crosstalk , we investigated the effect of SPP86 on the proliferation of MCF7 cells .
+Estrogen deprived and serum starved cells were cultured in the presence of 1 ng @/@ ml β @- estradiol ( E2 ) or 10 ng @/@ ml GDNF alone and in combination in the presence of 1 μM SPP86 for 7 days .
 SPP86 effectively inhibited E2 and @/@ or GDNF @- induced proliferation ( <i> p </i> < 0.05 ) ( Figure <xref ref-type="fig" rid="Fig5"> 5 </xref>@ A ) .
 In contrast , SPP86 did not inhibit proliferation when MCF7 cells were co-exposed to 1 ng @/@ ml E2 and 5 ng @/@ insulin under similar conditions ( Figure <xref ref-type="fig" rid="Fig5"> 5 </xref>@ B ) .
 We next compared the effect of SPP86 and tamoxifen on the proliferation of MCF7 cells .
-Estrogen deprived and serum starved cells were cultured in the presence of 1 ng @/@ ml β- estradiol ( E2 ) and 10 ng @/@ ml GDNF with increasing doses of either SPP86 or tamoxifen , in medium containing 1 ng @/@ ml β- estradiol ( E2 ) and 10 ng @/@ ml GDNF and incubated for 7 days .
+Estrogen deprived and serum starved cells were cultured in the presence of 1 ng @/@ ml β @- estradiol ( E2 ) and 10 ng @/@ ml GDNF with increasing doses of either SPP86 or tamoxifen , in medium containing 1 ng @/@ ml β @- estradiol ( E2 ) and 10 ng @/@ ml GDNF and incubated for 7 days .
 In these experiments , SPP86 and tamoxifen inhibited proliferation to a similar degree with IC @<sub> 50 </sub> values of 1.0 and 1.4 μM respectively ( Figure <xref ref-type="fig" rid="Fig5"> 5 </xref>@ C ) .
 Constitutive activation of the ERK pathway in melanoma and skin melanocytes in Grey horses ( PMID : 25413220 )
 <sec-title level="1" sec-area="abstract"> Results </sec-title>

v1.3.9/test/tok-challenge.tok-ref

@@ -0,0 +1,17 @@
+::raw ZO-1 , ZO-2 , and ZO-3 , which contain three PDZ domains ( PDZ1 to -@ 3 ) , are concentrated at tight junctions ( TJs ) in epithelial cells .
+::raw TJ strands are mainly composed of two distinct types of four @-@ transmembrane proteins , occludin , and claudins , between which occludin was reported to directly bind to ZO-1 @/@ ZO-2 @/@ ZO-3 .
+::raw <sup> * </sup> , <em> p </em> &lt; 0.05 , <em> n </em> = 3 by <em> t </em> test .
+::raw <sup> * </sup> , <em> p </em> &lt; 0.05 to control IgG , <em> n </em> = 3 by <em> t </em> test .
+::raw <sup> * </sup> , <em> p </em> &lt; 0.05 to DMSO , <em> n </em> = 3 by <em> t </em> test .
+::raw Interestingly , most of the claudin family members end in YV at their COOH termini , which are good candidates for the binding partners for PDZ domains ( see <xref ref-type="fig" rid="F5"> Fig. 5 </xref> B ) .
+::raw As shown in <xref ref-type="fig" rid="F1"> Fig. 1 </xref> , not only ZO-1 but also ZO-2 and ZO-3 were targeted to and concentrated at TJs of these well @-@ polarized occludin @-@ deficient epithelial cells .
+::raw We incubated 10 μM purified His <sub> 6 </sub> RbC <sup> 771 @–@ 928 </sup> with 10 μM of each purified E2F <sup> CM </sup> -@ DP <sup> CM </sup> heterodimer , precipitated it using Ni <sup> 2+ </sup> -@ NTA resin , eluted it with imidazole , and analyzed His <sub> 6 </sub> RbC <sup> 771 @–@ 928 </sup> bound proteins with SDS @-@ PAGE and Coomassie staining .
+::raw <a> Figure 1 </a> B shows that RbC <sup> 771 @–@ 928 </sup> bound all four E2F <sup> CM </sup> -@ DP <sup> CM </sup> heterodimers but did not bind the E2F1 transactivation domain used as a negative control .
+::raw Mutants displaying only minor perturbations ( |ΔΔG| &lt; 1 kcal mol @<sup> −1 </sup> ) are colored light gray , and mutants displaying increased K @<sub> d </sub> values are colored pink if the perturbations are moderate ( |ΔΔG| = 1 @–@ 1.5 kcal mol @<sup> −1 </sup> ) or red if the perturbations are dramatic ( |ΔΔG| &gt; 1.5 kcal mol @<sup> −1 </sup> ) .
+::raw Y4 and Y27 are located adjacent to the aromatic sandwich , with the Y4A mutation causing a moderate perturbation ( |ΔΔG| = 1.2 kcal mol @<sup> −1 </sup> ) and the Y27A mutation resulting in misfolded protein .
+::raw RbC is also required for high @-@ affinity binding to E2F @-@ DP complexes and for maximal repression of E2F @-@ responsive promoters ( <a> Hiebert , 1993 </a> , <a> Hiebert et al. , 1992 </a> and <a> Qin et al. , 1992 </a> ) .
+::raw The importance of the RbC domain is recapitulated in the Rb paralogs p107 and p130 , which have conserved sequences C-terminal to their pocket domains that are required for growth suppression , high @-@ affinity binding to their preferred E2F @-@ DP complexes , and maximal repression of E2F @-@ responsive promoters ( <a> Zhu et al. , 1995 </a> ) .
+::raw In addition , Drk , the <em> Drosophila </em> homologue of mammalian Grb2 and Mona @/@ Gads , binds two sequence motifs in the ‘ Daughter of Sevenless ’ ( Dos ) protein which strongly resemble the SLP-76 binding site of Grb2 and Mona @/@ Gads ( P–x <sub> 3 </sub> –R–x <sub> 2 </sub> –K–P ) ( <xref ref-type="bibr"> Feller et al. , 2002 </xref> ) .
+::raw Several mammalian signalling proteins also contain sequences with the P–x <sub> 3 </sub> –R–x <sub> 2 </sub> –K–P motif ( <xref ref-type="bibr"> Berry et al. , 2002 </xref> ; <xref ref-type="bibr"> Feller et al. , 2002 </xref> ) .
+::raw atvejį , bažnyčią .
+::raw 10 km² ; &#160;