bio-dbla-classifier 0.5.0 → 0.5.1
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- data/Gemfile +1 -0
- data/Gemfile.lock +2 -1
- data/README.rdoc +6 -2
- data/Rakefile +1 -1
- data/VERSION +1 -1
- data/bio-dbla-classifier.gemspec +4 -1
- data/lib/bio/sequence/aa/aa.rb +27 -11
- metadata +23 -12
data/Gemfile
CHANGED
data/Gemfile.lock
CHANGED
|
@@ -8,7 +8,7 @@ GEM
|
|
|
8
8
|
bundler (~> 1.0)
|
|
9
9
|
git (>= 1.2.5)
|
|
10
10
|
rake
|
|
11
|
-
rake (0.9.2)
|
|
11
|
+
rake (0.9.2.2)
|
|
12
12
|
rcov (0.9.10)
|
|
13
13
|
rspec (2.3.0)
|
|
14
14
|
rspec-core (~> 2.3.0)
|
|
@@ -26,5 +26,6 @@ DEPENDENCIES
|
|
|
26
26
|
bio (>= 1.4.2)
|
|
27
27
|
bundler (~> 1.1.rc.7)
|
|
28
28
|
jeweler (~> 1.6.4)
|
|
29
|
+
rake (~> 0.9.2.2)
|
|
29
30
|
rcov
|
|
30
31
|
rspec (~> 2.3.0)
|
data/README.rdoc
CHANGED
|
@@ -44,9 +44,13 @@ install the gem. This will install the bioruby gem if it is not already installe
|
|
|
44
44
|
|
|
45
45
|
#get the length of the tag
|
|
46
46
|
puts dbl_seq.size #=> 115
|
|
47
|
-
|
|
47
|
+
|
|
48
|
+
#determine whether the tag is a var1
|
|
48
49
|
dbl_seq.is_var1? #=> false
|
|
49
50
|
|
|
51
|
+
#is this tag a groupA like sequence tag?
|
|
52
|
+
dbl_seq.is_groupA_like?
|
|
53
|
+
|
|
50
54
|
= Finding the Position Specific Polymorphic Blocks(PSPB)
|
|
51
55
|
|
|
52
56
|
The pspb methods take 2 arguments, an anchor position and a window length that
|
|
@@ -72,7 +76,7 @@ window length is 14
|
|
|
72
76
|
#for each entry in the file
|
|
73
77
|
Bio::FlatFile.open(seq_file).each do |entry|
|
|
74
78
|
tag = Bio::Sequence::AA.new(entry.seq)
|
|
75
|
-
puts "#{entry.definition},#{tag.
|
|
79
|
+
puts "#{entry.definition},#{tag.cyspolv_group},#{tag.dsid},#{tag.bs_group},"
|
|
76
80
|
end
|
|
77
81
|
|
|
78
82
|
= Copyright
|
data/Rakefile
CHANGED
data/VERSION
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
0.5.
|
|
1
|
+
0.5.1
|
data/bio-dbla-classifier.gemspec
CHANGED
|
@@ -5,7 +5,7 @@
|
|
|
5
5
|
|
|
6
6
|
Gem::Specification.new do |s|
|
|
7
7
|
s.name = "bio-dbla-classifier"
|
|
8
|
-
s.version = "0.5.
|
|
8
|
+
s.version = "0.5.1"
|
|
9
9
|
|
|
10
10
|
s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
|
|
11
11
|
s.authors = ["George Githinji"]
|
|
@@ -47,12 +47,14 @@ Gem::Specification.new do |s|
|
|
|
47
47
|
s.add_development_dependency(%q<bundler>, ["~> 1.1.rc.7"])
|
|
48
48
|
s.add_development_dependency(%q<jeweler>, ["~> 1.6.4"])
|
|
49
49
|
s.add_development_dependency(%q<rcov>, [">= 0"])
|
|
50
|
+
s.add_development_dependency(%q<rake>, ["~> 0.9.2.2"])
|
|
50
51
|
else
|
|
51
52
|
s.add_dependency(%q<bio>, [">= 1.4.2"])
|
|
52
53
|
s.add_dependency(%q<rspec>, ["~> 2.3.0"])
|
|
53
54
|
s.add_dependency(%q<bundler>, ["~> 1.1.rc.7"])
|
|
54
55
|
s.add_dependency(%q<jeweler>, ["~> 1.6.4"])
|
|
55
56
|
s.add_dependency(%q<rcov>, [">= 0"])
|
|
57
|
+
s.add_dependency(%q<rake>, ["~> 0.9.2.2"])
|
|
56
58
|
end
|
|
57
59
|
else
|
|
58
60
|
s.add_dependency(%q<bio>, [">= 1.4.2"])
|
|
@@ -60,6 +62,7 @@ Gem::Specification.new do |s|
|
|
|
60
62
|
s.add_dependency(%q<bundler>, ["~> 1.1.rc.7"])
|
|
61
63
|
s.add_dependency(%q<jeweler>, ["~> 1.6.4"])
|
|
62
64
|
s.add_dependency(%q<rcov>, [">= 0"])
|
|
65
|
+
s.add_dependency(%q<rake>, ["~> 0.9.2.2"])
|
|
63
66
|
end
|
|
64
67
|
end
|
|
65
68
|
|
data/lib/bio/sequence/aa/aa.rb
CHANGED
|
@@ -5,10 +5,20 @@ class Bio::Sequence::AA
|
|
|
5
5
|
end
|
|
6
6
|
|
|
7
7
|
def start_motif
|
|
8
|
-
|
|
8
|
+
warn "[DEPRECATION] 'start_motif' is deprecated. Please use `n_terminal_motif` instead."
|
|
9
|
+
n_terminal_motif
|
|
9
10
|
end
|
|
10
11
|
|
|
11
12
|
def end_motif
|
|
13
|
+
warn "[DEPRECATION] 'end_motif' is deprecated. Please use `c_terminal_motif` instead."
|
|
14
|
+
c_terminal_motif
|
|
15
|
+
end
|
|
16
|
+
|
|
17
|
+
def n_terminal_motif
|
|
18
|
+
self[0,5]
|
|
19
|
+
end
|
|
20
|
+
|
|
21
|
+
def c_terminal_motif
|
|
12
22
|
self[-5,self.length]
|
|
13
23
|
end
|
|
14
24
|
|
|
@@ -76,6 +86,19 @@ class Bio::Sequence::AA
|
|
|
76
86
|
group
|
|
77
87
|
end
|
|
78
88
|
|
|
89
|
+
#return the block sharing group
|
|
90
|
+
def bs_group
|
|
91
|
+
case
|
|
92
|
+
when self =~ /(?:R(?:HYADHDKSGNYYK|NENNNLGKLSNEQ|I(?:RHYDDGSGNY(?:SK|YK)|THYN(?:GVSGNCVK|DGSGNYVK))|E(?:RYKDLKDVEIDD|HYKEVKNGNY(?:YK|IK)|KYKDLKD(?:VEIDD|LPIDD))|VKETYKDDPNYYK|KNNSSLRKLTNEQ)|G(?:G(?:RGRK(?:KLEDNL(?:IE|KE)|QLEENLQK)|GGRKKLEDNLKE)|I(?:N(?:D(?:Y(?:ND(?:GSGNYFK|ISGNYYK)|D(?:RDGPE(?:HYK|YYK)|GDGPEYYK))|CDRDGPEYYK)|AYNDGSENYYK)|IDYDHDGPHYYK)|P(?:SQEK(?:IKLEENLK|KKLEENLK)|KQEKKELEENLK)|K(?:NYPDDGSGN(?:YYK|FYK)|K(?:YYND(?:G(?:SGNYYK|TGNYVK)|ETGNYYK)|KEKEKIYGNIE)|QYYNDENGNYYK)|QTYPDDGSGNYYK)|MESNANLKKHTLER|S(?:STNTQCRCATNDV|HY(?:TDTHGSIDYDK|ED(?:GDKSGNYYK|KDKSGNYIK)|ADHDKSGNY(?:YK|LK))|NKEKEKIENSLQN|DYKDDD(?:GSGNYYK|IDGNYY(?:K|Q))|E(?:GKCGHKETERDL|KVEYGLRKLFKK)|VQERYGNDPNFFQ|KI(?:NDYDGDPNYYK|TDYDNDPNYYK)|QGQCGRNENNGYP|F(?:S(?:SEYCGH(?:RQGS(?:V|A)|GDNEV|EQGNV)|N(?:GQCGHRDENV|SKCGHGEHEV|DYCGH(?:RQGSV|NENKV)|PKCGH(?:G(?:DNEV|EHEV)|NENKV|EQGNV)|EYCGHRQGSV)|D(?:RKCGHYEGAP|HKCGH(?:GDKDV|YEGAP|DENAP)))|TN(?:GQCGHNEENV|PKCG(?:RGDNEV|H(?:G(?:DNEV|EHEV)|DENKV)))|ADAYCGRGDENV)|LILPYSKCGRDTD)|H(?:HYKDDDISGNYSK|SKEKEKLQTNLKN|N(?:H(?:IKKPLLENLEQ|KKKPLL(?:DNLEK|ENLEQ))|NKKKALLDNLEK|QKKINLEKSLHR)|EQG(?:YNKLEAI(?:SKT|LKT)|NNKLEA(?:RLKT|ILKT))|QQRK(?:RKLEENLRN|GKLEENLRN))|C(?:RAP(?:N(?:GANYFRKGL|EANYFKNVA)|KNAHYFIKSS|QKANYFKNVA)|G(?:TGENDTYFKNSS|A(?:G(?:EKDTYF(?:TYS(?:N|K)|VQLD)|A(?:RDEYFIKPS|KDTYFTYSK))|TMNDIFSKNIG|LPKSAY(?:VLQSE|F(?:MQLE|LQSE))))|S(?:VPYEAYYFTYKS|A(?:G(?:PKDTYFIKSG|QKDTYFIKPN)|D(?:GSE(?:DYFIKSS|EYF(?:IQSE|KKQS))|DSEDYFI(?:RSE|QSE))|P(?:RDA(?:DYFIKNS|QYFIKSS)|GD(?:VNYFRK(?:GL|IS|ES|FS)|AKYVK(?:YFP|NFP))|HNAQYVKYVP|Y(?:GANYYRKYS|HPGYFRQSK|YADYF(?:RKGS|K(?:SVA|KK(?:S|P)))|NAHYFIKSS|CADYFKKKS|DANYVRRKS|EA(?:YYFTYKS|QYFIKSS)|KSQYFIKSS)|D(?:YAKYFRQTC|NAKY(?:VKYFP|FKPPK)))|QNNEVYFINSE))|Y(?:IPY(?:YVNYFK(?:NIS|DIS|K(?:T(?:S|P)|KS))|CVNYFKNIS)|APNNANYFIGSG)|NAP(?:GD(?:VHYFRKDP|AHYFRKDP)|Y(?:DANY(?:YR(?:KYS|QTC)|VRRKS|FRKTS)|EAQY(?:YIKSS|FIK(?:SS|PS))|K(?:SRYF(?:M(?:HSE|QSE)|IQSE)|A(?:QYYIKSS|WYFMHSE)))|NISGYFMQS(?:G|E)|D(?:NVNYFRKYS|KAEYFVYKS))|TAP(?:YGANYYRKYS|D(?:NVNYFRKYS|KANYFIYKS))|IAPRDAHYFLKSS|D(?:TEESDTYFKQSS|A(?:SYKSGYFMQSE|P(?:RDA(?:HYFLKSS|NFFIKNS|DYF(?:RKGS|KNVA))|YKSRYF(?:MQSE|IQSE)|KDANYFIGSG|QKVDYFRK(?:GS|IS))))|EA(?:SKNANFFIK(?:NS|DS)|P(?:GDAHYFRKGP|ENAYIIKRRI|KDANYFIGSG|QKVDYFRKG(?:S|L)))|VA(?:GEGNTYFIQLD|PENAYFRKTEA)|KA(?:S(?:RNA(?:HYFLKSS|NYFRK(?:IS|AL))|KNANFFIKNS)|P(?:GDVN(?:YFRKIS|FFIKNS)|NGANYFRKKS|TGA(?:HYFLKSS|DYF(?:VYKP|KKKS))|PKVDYSRNIS|EDADYFRKGS|K(?:GANYFRKES|DA(?:HYFLKSS|N(?:YFIGSG|FFIKIS)|DYFRKGS))|Q(?:GANYFRNIS|SVHYFIKTS|DANYF(?:R(?:N(?:IS|VS)|K(?:GL|IS))|TKES)|K(?:VDYFRKGS|ANYFRKGS)))|KEGDIYSKTTD)|FA(?:HNTEEYFIKSE|DGSEEYFI(?:KSS|QS(?:S|E)))|AAR(?:GNDLYSKNIG|YHPGYFKKSD))|Y(?:NERDR(?:DKKRKLQE|EKKRKLQD|AQKKKLQD)|DEKEKNRRKQLEN|KAP(?:RKA(?:NYFIYKS|DYFRNIS)|KDAHYFLKSS|QDANYFRNVS))|N(?:RKEKGKLQTNLKN|GDYK(?:EKVSNNL(?:RA|KT)|KKVSNNLKT)|HYKDD(?:NGS(?:GNYYK|ENYYK)|D(?:GSGNYYK|ISGNY(?:SK|YK)))|SDDKVE(?:NGLKKVF|KGLREVF)|Y(?:Y(?:NNTGNN(?:V(?:NYAK|DYVK)|ANYAK)|ADGDKSGNYYK)|NYDEDGSGNY(?:YK|VK))|N(?:HDNVE(?:NGL(?:REVF|KAVF)|KGLK(?:KVF|AVF))|D(?:ND(?:RVKKEKLQN|K(?:IKK(?:GKLRG|EKL(?:RG|QE))|VKKEKL(?:RG|QN)))|DDKIKKGKLRG|V(?:EKGL(?:DVVFKK|KVVFKK)|VKGLDVVFKK))|E(?:SE(?:IKRKEKLRG|KKKREELQG)|TDKEQKVKLEK|KDM(?:REKQKLQS|TEKQKLQS))|VDAVQEGLKVVF|KE(?:NEKLQENLKR|KEKIEKSLQN))|TVDK(?:IHEGLKVVF|VHEGLKVVF)|D(?:NVE(?:NGLREVFKK|KGLK(?:KVFDK|AVFRK))|DVEKGLKIVFEK|EDDVEKGLKIVF|K(?:D(?:YVENGLKKVF|A(?:V(?:RHGLKVVF|QKGLRAVF)|AQKVLRTVF))|EKDQRKKLDE(?:N|I)|VE(?:NGL(?:REVFKK|KKVFDK)|KGL(?:REVF(?:RK|KK)|QVVFGK)))|QD(?:DVEKGLKIVF|EVWNGLRSVF)|ADKV(?:EKGLQVVF|QKGLQVVF))|P(?:E(?:DKVHEGLKVVF|VE(?:NGLREVFNK|KGLKAVFRK))|QDKVQ(?:EGLK(?:NVF|VVF)|KGLREVF))|E(?:MVEIGLKKVFKK|HYKEVKNGNYVK|NVEKGL(?:K(?:IVFEK|KVFDK)|QVVFGK)|D(?:DKVQKGLQVVF|VEKGLKVVF(?:KK|QK))|EDAVQKGL(?:RAVF|K(?:VVF|KVF))|K(?:DAVQNGLKKVF|VE(?:YGLRKLFKK|IGLKKVF(?:DK|EK|KK)))|QDEVWKGLRDVF)|VH(?:YK(?:DDGS(?:GNYYK|ENYYK)|EVKNGNYVK)|DKVE(?:RGLREVF|TGLREVF|KGL(?:REVF|QVV(?:F|L))))|K(?:HDNIEKGLREVF|N(?:NVPL(?:HNLSLDK|DKLSLDK)|K(?:SPLDKLSLEQ|PPLDKLSVDK)|VEIGLK(?:NVFKN|KVFDK))|DDK(?:IEKSLRAIF|V(?:EKGLRAIF|QKGL(?:RAVF|KAVF|QVVF)))|Q(?:RKKILQEKLEN|EKEKREKLDEN))|FSNPKCGH(?:DEGIV|KQGNV)|QEDKVQEGLKVVF|LILTHPKCGHDTD)|I(?:SYYNADEKGNFYK|HNYDDNGSGNYYK|E(?:TRY(?:GSDTTNYYQ|ENDGPNYYQ)|ARYKKDDDNYYQ)|VSFDQCGHND(?:MDV|VDV)|KNDKTLNNLSNGQ|FSNEHCGHKQGSV)|T(?:SEGQCGHNDKMRP|HYADEDGS(?:GNYVK|ENYYK)|D(?:N(?:D(?:EVW(?:TGLRSVF|KGL(?:RSVF|GSVF))|AVQKGLRAVF)|VE(?:NGLREVFKK|KGLRAVFGK))|K(?:D(?:YVENGLK(?:KVF|AVF)|DVENGLREVF|EV(?:KEGLKVVF|WKGLRAVF)|AVQKGLRAVF)|VENGLK(?:EVFDK|KVFD(?:N|K))))|E(?:GYCGRNENNGYP|TLYKDEEGNYLK|KDDVEKGLKIVF)|V(?:S(?:SNKCGHNDMDV|NAKRREGDENP|FDQCGHNDM(?:HV|DV))|K(?:GTYKDDPYYYK|ETYKDDPNYYK))|F(?:S(?:GYWCGHYEGAP|NDYCGHGEHEV)|TYTKCGHDENKV))|D(?:SRTDKLEENLRKI|NNSDKLRDLSVDK|DDVEKGLKIVFEK|K(?:GEKKKLEKNLKD|NRGKLGALSLDD))|P(?:SY(?:IKCGHNNKDDP|LKCGHNNKDDP)|HYTNDRGLADYVK)|E(?:HY(?:EDVDGSGNYLK|KDVDGSGNYYK)|YY(?:NDTNNKINYVK|EDKDPDKNYYQ)|KNY(?:YNDGTGNYYK|PDDGSGNYSK)|FT(?:GGYCGRDETDV|SGYCGRNETNV))|V(?:NGNDKLESNLKKI|KAHY(?:KKDAPYYYK|QKDAPNYYK)|FSNRQCGHYED(?:VP|AP))|K(?:RYYNDDTD(?:NN(?:FYQ|LYQ)|DNFYQ)|GINDYDGDPNYYK|SYY(?:NAD(?:GEGNFYK|EKGNYYK)|DADEKGNYYK)|HY(?:TDTHGSIDYDK|A(?:H(?:GDGSGNY(?:SK|YK|LK)|DDGS(?:GNYYK|VNYYK))|DEDGSGNYYK))|YYNDTNNKINYVK|N(?:Y(?:YNPD(?:G(?:SGNYYK|AGNYYK)|EAGNYYK)|NYD(?:EDGPEYYK|KDGPEYYK))|NDRTLNNLSIGQ|DYNPDGSGNY(?:YK|FK)|E(?:SE(?:IKRKEKLQR|K(?:RTKEKLQG|NTKKKLQG))|NTDLNKLTTEK)|KNTKLSTLTLEK|AYPDDG(?:SGNY(?:YK|FK)|FGNYYK))|T(?:SN(?:SN(?:MDTLSLEQ|LKELSLDK)|TNMNTLSLDK)|IYADLKDVEIDD)|I(?:NDYDGDGPEYYK|THYDDISGNYYK|KDYDGDGPEYYK)|D(?:HY(?:KDEKDGNFFQ|QDDGTGNYYK)|YYNADEKGNYYK)|PHYKDDGFGNYYK|E(?:YYQDDGTGNYYK|ISDYDNDPNYYK|EYGDLKDVPIDD|KYGDLKDVPIDD)|V(?:HYKENKDGNY(?:YK|VK)|KY(?:PDL(?:ND(?:IEIDD|VEIDD)|KDLQIDD)|QDLKDVEIDD))|K(?:HY(?:ENDTDKNYYQ|KKDEDPNYYK)|VYPEDVTGNY(?:YK|FK)|KKKGLSELSTEK)|FS(?:SDRCGHNEGDP|ERKCGH(?:NEGSP|DENAP))|Q(?:HYK(?:DDGSVNYYK|EDKDENYYK)|NNKKLKDLTDKH)|A(?:RYKDRKDPNYYK|K(?:YEDLKTLPIDD|ERYKDIKNYYQ)))|F(?:S(?:S(?:SGPCGRDEAPV|HGKCGHNEGAP|DRCGHNNNDGP|E(?:GKCGHKEGTV|YCGHYKNGDP)|QGQCGHTEGTV)|N(?:RGPCGRNETDV|S(?:G(?:TRGRKELTV|PCGRKELTV|KCGGKEAPV)|KCGHHNNDGP)|N(?:GPCGRNETDV|KCGHSNGGDP)|PKCGHSNGGDP|E(?:HCGH(?:HNNDDP|YKNGDP)|YCGH(?:YKNGDP|KKNEDP)))|DNG(?:HCGRNETNV|PCGRKELIV))|T(?:RQGYCGHSETNV|GGGQC(?:RRNDNSV|GRNETDV)|S(?:HGKCG(?:RNETNV|HSEGAP)|IGKCGHNKGSV|EG(?:RCGHSETNV|KCG(?:RNETNV|HNDNRV)|Q(?:RGHSETNV|C(?:RRNDNSV|GH(?:SETNV|NDKSV|DENKV))))|VGYCGHNKG(?:SV|IR)|QG(?:YCG(?:RKE(?:LTV|APV)|HSETNV)|QCG(?:RNERNV|H(?:SETNV|TEGTV|KEGTV)))|AGKCR(?:RNDNSV|HNDNSV))|NDGKCG(?:RYEGAP|H(?:YE(?:GAP|NNI|D(?:NV|AP))|TEGTV))|TEGYCGR(?:NEGAP|DEGAP)|D(?:GHCGRTQEGHV|IGKCG(?:GKEAPV|H(?:GDKDV|N(?:EGAP|KGSV)|KQGNV))|DGKCGHYEGAP))|ENAGKC(?:RRNDNKV|GHNDNRV)|WDRKCGHSN(?:GGDP|EGA(?:P|L))|LYPKCGHNNKNDL)|Q(?:GIIDYDNDPNYYK|NYY(?:KDDPKKNYYK|ADDGSGNY(?:SK|VK))|ISDY(?:TGDHPNYYK|DGDGPEYYK)|K(?:SDSSLQRLSIEK|HYEDDGSGNYYK|NNSALKKLTDKQ|IY(?:EDINNLPIDD|KDLNNLPIDD))|QNNNTLENLTDKQ)|L(?:K(?:TRYKKDDDNYYQ|KHYQKDAPNYYK)|F(?:S(?:N(?:RQCGH(?:GEHEV|NEGAP|DENKV|EQGNV)|SKCGH(?:RQGNV|DE(?:SKV|NKV)|EQGNV)|YKCGHYEDAP|DYCGHKQGNV|PKCGH(?:EQG(?:NV|TV)|KQGKV)|AYCGHYEGSP)|D(?:GHCGNKDGTV|HKCGHEESRV|YKCGHYE(?:GSP|DAP)))|DYNCGHHKDNNV|WDRKCGHDERNV)|Q(?:TRYTNDGDNYFK|ERYNDPKGDFFQ|ARYKKDGDDFFK)|WNDKCGHHVDKDV|LFSNYKCGHYEGS)|A(?:RYKKDEEDGNYY(?:K|Q)|VSSNKCGHNDMNV|KNDYTGDHPNYYK|LKHYKDDTKNYYQ))/
|
|
93
|
+
block_sharing = 1
|
|
94
|
+
when self =~ /(?:RE(?:PGKQHLEERLER|KGKSRLEARLKT)|G(?:G(?:HYKNCHCIGGDV|TYKNCRCASGNV)|P(?:NQEKKLLENKLK|DQEKKKLEENLR)|ANAIKAGDNVSIV)|S(?:FTNGQCGRDGENV|WYPKCGHHVKQDV|AKEHYQDTENYYK)|H(?:NRRKEKLETRLEE|E(?:PG(?:IQ(?:HLEKRLES|YLEKRLES)|KQHL(?:GERLEQ|EERLE(?:R|Q)))|QG(?:YNRLEARLKT|NNRLEARLKT|INRLEARLKT))|QQRKHLLEKRLET)|C(?:RAEEK(?:GTYFKNRE|D(?:TYFKNRE|IYSKTTD)|EIYSKTTD)|G(?:TEDKDTYFIKSG|VEENAKYFRESS|A(?:GMKDIYSKTMN|SEDAKYKVIGP|T(?:MNDIFSKNIR|VDDI(?:SSKNIR|FSKNIR))|P(?:SDAQYFRNTC|KEAKYFRKTA)))|HAPPDAQYTKKGP|NA(?:P(?:TGADYFVYKP|KDANYFEYNS)|WGNTYFRKTCS)|DAG(?:QKDTYFKQSS|AADEYFKKSG)|EA(?:GTSDKYFRKTA|KSDDKYNVIGP)|KA(?:NDDAEYFRKKD|P(?:DKANYF(?:EPPK|KPPK)|EEDHYFKPAQ)|EVDDIYSKTAN|K(?:EGDIYSKT(?:MN|AN)|KG(?:GIYSKTMN|DIYSKTMN))))|Y(?:SQKYKDEKSKLEE|NE(?:TDKVQKAILQQ|KDQEEKRKLQE))|N(?:YYEDNDTDKNYYQ|NNAAKLSELSTAQ|THESAQRKKLEEN|DEEKKKRDELEKN|KN(?:NPPLYKLSLEK|KSPLDKLSLDK))|T(?:RYKKDDEDGNFFQ|PTQGKCHCIDGT(?:N|V)|L(?:FDYKCGHDENAP|WNEKCGHGDYNL))|I(?:EHYKDDPEENFYE|KS(?:NYNDSEGNYFK|QYDDNEGNYFK)|LFDYKCAHDNDKV)|D(?:RKEKVKLEENLKN|KGEKKKLEENLKN|Q(?:ERK(?:HLLEKRLET|QHLEKRLET)|QEK(?:LYLENNLKK|AKLENNLKR))|AKKHYGDDENYYK)|P(?:NKCRCEDANADQV|CSVQKCTCINGDP)|E(?:SN(?:M(?:GQCRCFSGDP|VQCRCFSGDP)|KGQCRCFSGDP)|THGYCRCVNRVDV|F(?:SGGKCGHKDNNV|T(?:GGQCGRDGENV|DGHCGH(?:RQGNV|NEENV))))|VKDRYQNDGPDFFK|K(?:S(?:SYNDDGTGNYFK|YYKNDNDRNYFK)|N(?:HYNDTSKNYYK|NNN(?:ELNNLSLDK|KLSNLSTKE))|D(?:HYKGDEANNYFQ|NNTKLNDLSIQE)|K(?:TNPALKSFTNEE|KLEENLRNIFKN)|FSNPKCGHNEGSP)|F(?:SNDQCGHNN(?:RGDP|GGAP)|WYPKC(?:GHHV(?:RQDV|KQ(?:DV|EV))|SHHVKQDV))|Q(?:RNNIKLQ(?:NIPLHE|TLTLHQ)|N(?:NNTKLQNIPLHE|KNENLKSLSLDK)|KENGDINTLKPEE)|L(?:FYYKCGHYVYKDV|WN(?:Y(?:NCGHHVN(?:RDV|QDV)|KCGHHVNQDV)|DKCGHHVKQDV))|ARDHYNDTSGNYYQ)/
|
|
95
|
+
block_sharing = 2
|
|
96
|
+
else
|
|
97
|
+
block_sharing = 0
|
|
98
|
+
end
|
|
99
|
+
block_sharing
|
|
100
|
+
end
|
|
101
|
+
|
|
79
102
|
def is_var1_cp1?
|
|
80
103
|
return true if cyspolv_group == 1 && self =~ /NVHDKVEKGLREVF|NVHDKVETGLREVF/i
|
|
81
104
|
end
|
|
@@ -88,16 +111,9 @@ class Bio::Sequence::AA
|
|
|
88
111
|
return true if is_var1_cp1? || is_var1_cp2?
|
|
89
112
|
end
|
|
90
113
|
|
|
91
|
-
|
|
92
|
-
|
|
93
|
-
|
|
94
|
-
block_sharing = 1
|
|
95
|
-
when self =~ /(?:RE(?:PGKQHLEERLER|KGKSRLEARLKT)|G(?:G(?:HYKNCHCIGGDV|TYKNCRCASGNV)|P(?:NQEKKLLENKLK|DQEKKKLEENLR)|ANAIKAGDNVSIV)|S(?:FTNGQCGRDGENV|WYPKCGHHVKQDV|AKEHYQDTENYYK)|H(?:NRRKEKLETRLEE|E(?:PG(?:IQ(?:HLEKRLES|YLEKRLES)|KQHL(?:GERLEQ|EERLE(?:R|Q)))|QG(?:YNRLEARLKT|NNRLEARLKT|INRLEARLKT))|QQRKHLLEKRLET)|C(?:RAEEK(?:GTYFKNRE|D(?:TYFKNRE|IYSKTTD)|EIYSKTTD)|G(?:TEDKDTYFIKSG|VEENAKYFRESS|A(?:GMKDIYSKTMN|SEDAKYKVIGP|T(?:MNDIFSKNIR|VDDI(?:SSKNIR|FSKNIR))|P(?:SDAQYFRNTC|KEAKYFRKTA)))|HAPPDAQYTKKGP|NA(?:P(?:TGADYFVYKP|KDANYFEYNS)|WGNTYFRKTCS)|DAG(?:QKDTYFKQSS|AADEYFKKSG)|EA(?:GTSDKYFRKTA|KSDDKYNVIGP)|KA(?:NDDAEYFRKKD|P(?:DKANYF(?:EPPK|KPPK)|EEDHYFKPAQ)|EVDDIYSKTAN|K(?:EGDIYSKT(?:MN|AN)|KG(?:GIYSKTMN|DIYSKTMN))))|Y(?:SQKYKDEKSKLEE|NE(?:TDKVQKAILQQ|KDQEEKRKLQE))|N(?:YYEDNDTDKNYYQ|NNAAKLSELSTAQ|THESAQRKKLEEN|DEEKKKRDELEKN|KN(?:NPPLYKLSLEK|KSPLDKLSLDK))|T(?:RYKKDDEDGNFFQ|PTQGKCHCIDGT(?:N|V)|L(?:FDYKCGHDENAP|WNEKCGHGDYNL))|I(?:EHYKDDPEENFYE|KS(?:NYNDSEGNYFK|QYDDNEGNYFK)|LFDYKCAHDNDKV)|D(?:RKEKVKLEENLKN|KGEKKKLEENLKN|Q(?:ERK(?:HLLEKRLET|QHLEKRLET)|QEK(?:LYLENNLKK|AKLENNLKR))|AKKHYGDDENYYK)|P(?:NKCRCEDANADQV|CSVQKCTCINGDP)|E(?:SN(?:M(?:GQCRCFSGDP|VQCRCFSGDP)|KGQCRCFSGDP)|THGYCRCVNRVDV|F(?:SGGKCGHKDNNV|T(?:GGQCGRDGENV|DGHCGH(?:RQGNV|NEENV))))|VKDRYQNDGPDFFK|K(?:S(?:SYNDDGTGNYFK|YYKNDNDRNYFK)|N(?:HYNDTSKNYYK|NNN(?:ELNNLSLDK|KLSNLSTKE))|D(?:HYKGDEANNYFQ|NNTKLNDLSIQE)|K(?:TNPALKSFTNEE|KLEENLRNIFKN)|FSNPKCGHNEGSP)|F(?:SNDQCGHNN(?:RGDP|GGAP)|WYPKC(?:GHHV(?:RQDV|KQ(?:DV|EV))|SHHVKQDV))|Q(?:RNNIKLQ(?:NIPLHE|TLTLHQ)|N(?:NNTKLQNIPLHE|KNENLKSLSLDK)|KENGDINTLKPEE)|L(?:FYYKCGHYVYKDV|WN(?:Y(?:NCGHHVN(?:RDV|QDV)|KCGHHVNQDV)|DKCGHHVKQDV))|ARDHYNDTSGNYYQ)/
|
|
96
|
-
block_sharing = 2
|
|
97
|
-
else
|
|
98
|
-
block_sharing = 0
|
|
99
|
-
end
|
|
100
|
-
block_sharing
|
|
114
|
+
# return var group A like tags. Group A like sequences are associated with disease severity.
|
|
115
|
+
def is_groupA_like?
|
|
116
|
+
return true if cys_count == 2 && bs_group == 1
|
|
101
117
|
end
|
|
102
118
|
|
|
103
119
|
#distict sequence identifier(DSID)
|
metadata
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
|
2
2
|
name: bio-dbla-classifier
|
|
3
3
|
version: !ruby/object:Gem::Version
|
|
4
|
-
version: 0.5.
|
|
4
|
+
version: 0.5.1
|
|
5
5
|
prerelease:
|
|
6
6
|
platform: ruby
|
|
7
7
|
authors:
|
|
@@ -13,7 +13,7 @@ date: 2012-02-20 00:00:00.000000000 Z
|
|
|
13
13
|
dependencies:
|
|
14
14
|
- !ruby/object:Gem::Dependency
|
|
15
15
|
name: bio
|
|
16
|
-
requirement: &
|
|
16
|
+
requirement: &2152826100 !ruby/object:Gem::Requirement
|
|
17
17
|
none: false
|
|
18
18
|
requirements:
|
|
19
19
|
- - ! '>='
|
|
@@ -21,10 +21,10 @@ dependencies:
|
|
|
21
21
|
version: 1.4.2
|
|
22
22
|
type: :runtime
|
|
23
23
|
prerelease: false
|
|
24
|
-
version_requirements: *
|
|
24
|
+
version_requirements: *2152826100
|
|
25
25
|
- !ruby/object:Gem::Dependency
|
|
26
26
|
name: rspec
|
|
27
|
-
requirement: &
|
|
27
|
+
requirement: &2152823580 !ruby/object:Gem::Requirement
|
|
28
28
|
none: false
|
|
29
29
|
requirements:
|
|
30
30
|
- - ~>
|
|
@@ -32,10 +32,10 @@ dependencies:
|
|
|
32
32
|
version: 2.3.0
|
|
33
33
|
type: :development
|
|
34
34
|
prerelease: false
|
|
35
|
-
version_requirements: *
|
|
35
|
+
version_requirements: *2152823580
|
|
36
36
|
- !ruby/object:Gem::Dependency
|
|
37
37
|
name: bundler
|
|
38
|
-
requirement: &
|
|
38
|
+
requirement: &2152837840 !ruby/object:Gem::Requirement
|
|
39
39
|
none: false
|
|
40
40
|
requirements:
|
|
41
41
|
- - ~>
|
|
@@ -43,10 +43,10 @@ dependencies:
|
|
|
43
43
|
version: 1.1.rc.7
|
|
44
44
|
type: :development
|
|
45
45
|
prerelease: false
|
|
46
|
-
version_requirements: *
|
|
46
|
+
version_requirements: *2152837840
|
|
47
47
|
- !ruby/object:Gem::Dependency
|
|
48
48
|
name: jeweler
|
|
49
|
-
requirement: &
|
|
49
|
+
requirement: &2152835560 !ruby/object:Gem::Requirement
|
|
50
50
|
none: false
|
|
51
51
|
requirements:
|
|
52
52
|
- - ~>
|
|
@@ -54,10 +54,10 @@ dependencies:
|
|
|
54
54
|
version: 1.6.4
|
|
55
55
|
type: :development
|
|
56
56
|
prerelease: false
|
|
57
|
-
version_requirements: *
|
|
57
|
+
version_requirements: *2152835560
|
|
58
58
|
- !ruby/object:Gem::Dependency
|
|
59
59
|
name: rcov
|
|
60
|
-
requirement: &
|
|
60
|
+
requirement: &2152831480 !ruby/object:Gem::Requirement
|
|
61
61
|
none: false
|
|
62
62
|
requirements:
|
|
63
63
|
- - ! '>='
|
|
@@ -65,7 +65,18 @@ dependencies:
|
|
|
65
65
|
version: '0'
|
|
66
66
|
type: :development
|
|
67
67
|
prerelease: false
|
|
68
|
-
version_requirements: *
|
|
68
|
+
version_requirements: *2152831480
|
|
69
|
+
- !ruby/object:Gem::Dependency
|
|
70
|
+
name: rake
|
|
71
|
+
requirement: &2153076260 !ruby/object:Gem::Requirement
|
|
72
|
+
none: false
|
|
73
|
+
requirements:
|
|
74
|
+
- - ~>
|
|
75
|
+
- !ruby/object:Gem::Version
|
|
76
|
+
version: 0.9.2.2
|
|
77
|
+
type: :development
|
|
78
|
+
prerelease: false
|
|
79
|
+
version_requirements: *2153076260
|
|
69
80
|
description: Methods to classify and manipulate PfEMP1 DBL-alpha sequence tags
|
|
70
81
|
email: georgkam@gmail.com
|
|
71
82
|
executables: []
|
|
@@ -103,7 +114,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
|
|
|
103
114
|
version: '0'
|
|
104
115
|
segments:
|
|
105
116
|
- 0
|
|
106
|
-
hash:
|
|
117
|
+
hash: 762142858629813033
|
|
107
118
|
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
108
119
|
none: false
|
|
109
120
|
requirements:
|