UK Genetic Evaluation thread

[NS1]

Fatal Divergency on 4 UK Zoonotic Sequences

Fatal Divergency on 4 UK Zoonotic Sequences

http://pf11.blogspot.com/2011/01/fatal-divergency-on-4-uk-zoonotic.html

** Error on spacing prior to table with sequence name norming **

2011-01-07

Fatal Divergency on 4 UK Zoonotic Sequences

The UK Health Protection Agency released an analysis of the early severe wave today in Eurosurveillance. The requested citation is:

Ellis J, Galiano M, Pebody R, Lackenby A, Thompson C, Bermingham A, McLean E, Zhao H, Bolotin S, Dar O, Watson JM, Zambon M. Virological analysis of fatal influenza cases in the United Kingdom during the early wave of influenza in winter 2010/11. Euro Surveill. 2011;16(1)ii=19760.
Available online:
http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19760
Date of submission: 31 December 2010

No additional sequences appear to have been released with the paper though a very useful phylogenetic chart is provided as Figure 3: Phylogenetic relationship of full-length HA sequences of influenza A(H1N1)2009 viruses from fatal, severe and mild cases in the United Kingdom during 2010.

All four previously released UK sequences from 2010-12-20 are on the Ellis Figure3 <SUP>1</SUP> chart marked as fatality and all four were updated yesterday on GISAID as "Deceased". Two were recent and all four are divergent.

Fatal Sequences from Ellis Figure 3 In GISAID 2010-12-20 Deposit
<TABLE><TBODY><TR><TD>* UKWhiteChapel4880374_2M_2010_11_28_f</TD><TD>= A/England/4880374/2010</TD></TR>
<TR><TD>* UKCambridge118_4F_2010_11_19_f</TD><TD>= A/England/118/2010</TD></TR>
<TR><TD>* UKWhiteChapel4780352_5M_2010_10_26_f</TD><TD>= A/England/4780352/2010</TD></TR>
<TR><TD>* UKBirmingham3220137_44F_2010_08_07_f</TD><TD>= A/England/3220137/2010</TD></TR>
</TBODY></TABLE>

GeneWurx has annotated the Ellis Figure3 <SUP>1</SUP> phylogenetic chart with green boxes next to the four fatal GISAID 2010-12-20 deposited sequences and has proposed a 225G branch due to lack of data transparency.

• UK_2010_Phylo_ELLIS_Fig3new_4_GISAID_Notated.jpg
  
• GeneWurx_UK_December_Emerging_Genetics_v3.xls
  

Preliminary Comments on the Ellis Figure 3
(without the essential corroborating sequences)

The numbering system employed in our analyses requires adding 3 to the amino acid positions indicated in the UK chart.

Notice that the branch polymorphisms are not labeled from that top D97N (100N) upwardly.

The accumulation of 100N, 188T, 377K, 225G and 454N patterns onto the OzBrisbane209_51F_2010_08_09 sequence from the late winter in the Southern Hemisphere. Though parameters are adjustable for the phylogenetic chart, the top section may indicate one or more unmarked branches near the top right. Either publication of the sequences or the notation for the branching polymorphism would be very useful at this time. Are these strictly indicative of the high variability head and tail changes being seen frequently (under aa100, over aa499)?

Is it probable that no instance of change at any amino acid position from 156 to 159 has been elucidated from this wide inspection of the severe wave?

Alternate sub-clade:

By chart appearances, the accumulation of 128D onto the more established backgrounds with 100N and 377K (lower section) creates a fatality risk similar to more recognised severity markers. The bottom fatal sequence designated on the lower branch as A/England/4780352/2010 is identified as UKWhiteChapel4780352_2010_10_26_f in v3 of the GeneWurx Emerging Genetics spreadsheet within column AY and carries three additional silent (synonymous) changes at syn58C, syn131S and syn210S.

Hyper-Zoonotic Polymorphism Details

Please excuse the intermediate markings on these sequences. The following are rough lab notes, but divergence may be ascertained easily.

The common ground among these sequences and those in the database on either side of them temporally, from Cameroon (Sub-clade 1 & Sub-clade 2) and especially Iran, is the high quantity of polymorphisms that are novel or rare to pH1N1 and that also appear in zoonotic reservoirs, particularly H3N8, H5N1 and H7N7.

Even small genetic adjustments from animal vectors into human infections, especially H3N8, may create variant behaviour. This potential era of zoonotic sub-segment spillover merits as much investigation for severity (in combination) as does the recognised 225G.

. . . . UKWhiteChapel4880374_2M_2010_11_28_f (
. . . . . . . . 0A (gCC) [1918 (GCT), S7 (GCT), S5 (GCA)]
. . . . . . . . syn55L [S9, H5N1],
. . . . . . . . . . . . . . [Darwin47_2010_08_09 with syn529L,
. . . . . . . . . . . . . . Iran16273_2009_11_22 with 226R
. . . . . . . . . . . . . . NZ_Waikato2_2010_01_04 with 233H,
. . . . . . . . . . . . . . tkOntarioFAV117_1C_2009_12_07 mult domain matches, et al],
. . . . . . . . 188T [H6N1, H7N7],
. . . . . . . . syn338G [H3N8, H4, H5, H6, sw],
. . . . . . . . . . . . . . . [OzBrisbane209_51F_2010_08_09
. . . . . . . . . . . . . . . . . . . . with 156E & 225G,
. . . . . . . . . . . . . . . Arizona05_2010_05_11 with 0A,
. . . . . . . . . . . . . . . Swine Asia 2005 with 0A, et al]
. . . . . . . . 377K,
. . . . . . . . 454N [H7N3, H7N7, H9N2]
. . . . . . . . . . [Florida14_24M_2010_08_05
. . . . . . . . . . . . . . . . . . with 188T, 454N,
. . . . . . . . . . FL_Pen210_2009_11_10
. . . . . . . . . . . . . . . . . . with 225E,
. . . . . . . . . . SouthCarolina18_2009_09_16_VxX
. . . . . . . . . . . . . . . . . . with 159D, 224K, et al],
. . . . . . . . syn529L (CTt) [Unique to PF11]
. . . . . . . . . . . . . . . . . . . . [S7 (CTt), tn with syn338G (GGg)]
. . . . . . . . . . . . . . . . . . . . [PF11 32 Worldwide (CTa),
. . . . . . . . . . . . . . . . . . . . PF11 5 North America (tTG)],
. . . . . . . . . . . . . . . . . . . . [swThailandCU_CHK4_2009_01 (tTG)
. . . . . . . . . . . . . . . . . . . . . . . . . . with 0A, syn338G, 189T, 377G, syn451K, syn456L])

. . . . UKCambridge118_4F_2010_11_19_f (
. . . . . . . . syn118E tcatttgaaaggtttgaA,
. . . . . . . . 137T [MoldovaG170_2009,
. . . . . . . . . . . . . 41 sequences at GISAID],
. . . . . . . . syn163K,
. . . . . . . . 186P,
. . . . . . . . syn251L gaagcaactggaaatctG,
. . . . . . . . syn293L gctataaacaccagcctT,
. . . . . . . . syn363G [21 sequences at GISAID],
. . . . . . . . syn455Q caAttaaaaaacaatgcc ,
. . . . . . . . syn474C [H3N8],
. . . . . . . . 504G ttaaacagagaagaaatagGt,
. . . . . . . . 513V [1918, S5, tn] Gtttaccagattttggcgatc)

. . . . UKWhiteChapel4780352_5M_2010_10_26_f (
. . . . . . . . syn58C,
. . . . . . . . 100N,
. . . . . . . . 128D,
. . . . . . . . syn131S tcaaacaaaggtgtaacggca,
. . . . . . . . syn210S,
. . . . . . . . 377K)

. . . . UKBirmingham3220137_44F_2010_08_07_f (
. . . . . . . . #8A gcagttctgctatatacattt,
. . . . . . . . 175K aaagtcctcgtgctatgg,
. . . . . . . . 311E Gaaagcacaaaattgaga,
. . . . . . . . 377K,
. . . . . . . . syn385V gtCattgaaaagatgaat,
. . . . . . . . syn451K aagaacttatatgaaaaA,
. . . . . . . . syn454S agTcagttaaaaaacaatgcc,
. . . . . . . . syn494E,
. . . . . . . . 537G [S5] tccctgggggcaatcGgt)

1. Ellis J, Galiano M, Pebody R, Lackenby A, Thompson C, Bermingham A, McLean E, Zhao H, Bolotin S, Dar O, Watson JM, Zambon M. Virological analysis of fatal influenza cases in the United Kingdom during the early wave of influenza in winter 2010/11. Euro Surveill. 2011;16(1)ii=19760. Available online: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19760

 

[NS1]

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

The HPA released sequences today at GISAID. 225G only appears twice while instances of 221L and 224K enter the UK geography.

Multiple, hyper-zoonotic sub-clades are circulating with the well-discussed Australian 188T base in place.

Genetic Diversity is the norm.

 

[NS1]

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

A group of the UK sequences appear to be predecessors to Iran. The H3N8 homologies are in place in the UK and Iran while Iran then acquired the H7N7 syn411N. Multi-zoonotic.

Wisconsin08_2010_08_10 has 7 of the same SNPs as IranShahriar5336_2010_12_06. 5 of those 6 are from H3N8 animals.

. . . . IranShahriar5336_2010_12_06 (
. . . . . . . . 137T (aCA) [H3N8 donor aAT, aGT, aGC],
. . . . . . . . 186P,
. . . . . . . . syn297N [H3N8 gadwallRussiaAltai1325_2007_09],
. . . . . . . . . . . . . . . [H5N1],
. . . . . . . . . . . . . . . [Wisconsin08_2010_08_10
. . . . . . . . . . . . . . . . . . . . . . with 39R, syn78S, 137T, 186P, syn297N,
. . . . . . . . . . . . . . . . . . . . . . . . . . syn326S, syn346G, syn388K,
. . . . . . . . . . . . . . . . . . . . . . . . . . syn413K, 444K, syn474C,
. . . . . . . . . . . . . . . CalifVRDL2_2010_01_11
. . . . . . . . . . . . . . . . . . . . . . with syn121P & 502K,
. . . . . . . . . . . . . . . YaroslavlIIV196_2009_12_04_f
. . . . . . . . . . . . . . . . . . . . . . with syn159N, 225G,
. . . . . . . . . . . . . . . WiscD0128_2009_11_15
. . . . . . . . . . . . . . . . . . . . . . with syn343G, 377G, 471H,
. . . . . . . . . . . . . . . Brussels243_2009_11_09
. . . . . . . . . . . . . . . . . . . . . . with syn44L, syn159N & syn323N,
. . . . . . . . . . . . . . . Australia6_2009_07_18
. . . . . . . . . . . . . . . . . . . . . . with syn159N, 233H]
. . . . . . . . syn326S [Wisconsin08_2010_08_10
. . . . . . . . . . . . . . . . . . . . . . with 39R, syn78S, 137T, 186P, syn297N,
. . . . . . . . . . . . . . . . . . . . . . . . . . syn326S, syn346G, syn388K,
. . . . . . . . . . . . . . . . . . . . . . . . . . syn413K, 444K, syn474C,
. . . . . . . . . . . . . . . Zhongyuan1643_2009_11_16
. . . . . . . . . . . . . . . . . . . . . . with 208G],
. . . . . . . . syn383N [H3N8],
. . . . . . . . syn388K [H3N8, H5N1, S7],
. . . . . . . . . . . . . . . [OzVictoria508_2010_07_24
. . . . . . . . . . . . . . . . . . . . . . . with 238D,
. . . . . . . . . . . . . . . swIowa44837_1_2009_11_08_xL
. . . . . . . . . . . . . . . . . . . . . . . with 188R, 225N & 230I,
. . . . . . . . . . . . . . . Utah20_C2_2_2009_07_25_VxX
. . . . . . . . . . . . . . . . . . . . . . . with 159D & 227G, et al],
. . . . . . . . syn411N (AAc) [H7N3 & H7N7 donor ATc],
. . . . . . . . 444K (AAg) [H3N8 gAg],
. . . . . . . . syn474C [H3N8, Avian H1N1 2010],
. . . . . . . . . . . . . . . [Michigan10_2009_06_03 with 137T, 225N, et al])

 

[NS1]

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

Re: UK: Reports of Approximately 57 Deaths and 800+ Patients in intensive care (50 deaths confirmed by HPA as for Jan. 06 2011) due to influenza

158E in a mix found on 188T background. 159K with 128D.

As expected.

 

[NS1]

377G in UK Not Notated on HPA Chart

377G in UK Not Notated on HPA Chart

The 377G that is emergent in 2009 Human H5N1 from Dakahlia is also emergent in <?xml:namespace prefix = st1 ns = "urn:schemas-microsoft-comffice:smarttags" /><st1:country-region w:st="on">England</st1:country-region> and <st1:country-region w:st="on"><st1lace w:st="on">Iran</st1lace></st1:country-region>. A recent fatal case from England, UKEngland4500186_2010_11_f, carries the same coding for 377G combined with a 190Y only found elsewhere in 2010 Malmoe Sweden pH1N1 and in Avian H6N1.

The HPA chart of sequences did not notate the 377G branch though 5 English sequences, including this fatal one, appear on the chart carrying the H5N1 Human adaptation.

• UKEngland83_2010_10
• UKEngland87_2010_10
• UKEngland5500192_2010_10
• UKEngland119_2010_11
• UKEngland4500186_2010_11_f

<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>
Data indicates that this disease is rapidly diversifying by accumulating data from multiple animal reservoirs. The most recent <st1:country-region w:st="on"><st1lace w:st="on">UK</st1lace></st1:country-region> release is essentially divergent using hyper-zoonoses.

 

[NS1]

225G Only in 2 Sequences from Severe UK Wave

225G Only in 2 Sequences from Severe UK Wave

The UK Health Protection Agency released a group of 37 sequences at GISAID dated 2010-01-05 that appeared sometime after that date. Some of the sequences in this deposit also appear on the Figure 3 Phylogenetic Tree from the HPA Eurosurveilance paper requested to be cited as:
<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>
Ellis J, Galiano M, Pebody R, Lackenby A, Thompson C, Bermingham A, McLean E, Zhao H, Bolotin S, Dar O, Watson JM, Zambon M. Virological analysis of fatal influenza cases in the United Kingdom during the early wave of influenza in winter 2010/11. Euro Surveill. 2011;16(1)ii=19760.
Available online:
http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19760
Date of submission: 31 December 2010

The science community expected a full deposit to back the paper, but not all sequences noted on the phylogenetic table were in place at GISAIS as of this 2nd deposit.

The recent GISAID deposit added 4 of the fatalities noted on the phylogenetic tree to the 4 fatalities previously released on 2010-12-20. The previously released sequences were not marked at GISAID with a fatality notation, but now have been notated as "Deceased" as of yesterday. Oddly enough, on the 2010-01-05 deposit, none of the sequences notated as fatal on the phylogenetic tree were marked as to outcome at GISAID?

None of the sequences at the GISAID database that correspond to notated fatalities on the Eurosurveillance paper carry the severity marker of HA 225G.

225G is found on identical HA sequences noted as 2 severe cases.

• UKEngland4880378_2010_12
• UKEngland4940476_2010_12

The two sequences fall on a 100N, 188T, syn338G, 377K with 454N background. Appended to that background is a silent 179L previously found in US military sequences and as a regional marker in an alternate coding. A silent 448L completes the 225G sequences and is demonstrated in at least on fatal case, including in Russia.

. . . . UKEngland4940476_2010_12 (
. . . . . . . . 100N,
. . . . . . . . syn179L (CTg) [Regional Marker 2009 (tTA)]
. . . . . . . . . . . . . . . . . . . . [TexasAF2588_2009_10_04,
. . . . . . . . . . . . . . . . . . . . TexasJMS404_2010_01_08],
. . . . . . . . 188T,
. . . . . . . . 225G,
. . . . . . . . syn338G,
. . . . . . . . 377K,
. . . . . . . . syn448L [Yakutsk_EAV_2009_11_18,
. . . . . . . . . . . . . . . Yaroslavl_CHMV_2009_11_10_f with 224K & 225G mix],
. . . . . . . . 454N)

. . . . UKEngland4880378_2010_12 (
. . . . . . . . 100N,
. . . . . . . . syn179L,
. . . . . . . . 188T,
. . . . . . . . 225G,
. . . . . . . . syn338G,
. . . . . . . . 377K,
. . . . . . . . syn448L,
. . . . . . . . 454N)

 

[mixin]

Re: UK Genetic Evaluation thread

I ran all the HA nucleotides of the UK sequences with gs's program and then grouped them together in a way that looked right. Keep in mind that I'm no scientist. Maybe someone else will see different clades; please feel free to offer opinions.

The severe and fatal cases are marked; notice that about half of them appear on clade B.

Clade A: 451,598,658,1408,1464
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)

G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)
G451A,T598C,T658A,T858C,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A
G451A,T598C,G605A,T658A,A1172G,A1213G,G1266A,C1408T,C1464T,G1577T,G1630A(4)
G451A,T598C,G610T,T658A,A1172G,G1266A,C1408T,C1464T,G1577T,G1630A(4) F

G451A,G531A,T598C,T658A,C921T,A1131G,C1408T,C1464T(4)

Clade B: 340,658,1056,1171,1403,1408
G340A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,T1653C(4)
G340A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,T1653C(4)
G340A,G605C,T658A,T680C,T1056C,G1171A,G1403A,C1408T,T1653C(4) F
G340A,G605C,C612T,T658A,T1056C,G1171A,G1403A,C1408T,T1653C(4)

G340A,A579G,G605C,T658A,C704T,T1056C,G1171A,A1386G,G1403A,C1408T(4)
G340A,A579G,G605C,T658A,T1056C,G1171A,A1386G,G1403A,C1408T(4) S
G340A,A579G,G605C,T658A,T1056C,G1171A,A1386G,G1403A,C1408T(4) S
G340A,A579G,G605C,A611G,T658A,T1056C,G1171A,A1386G,G1403A,C1408T(4) F
G340A,A579G,G605C,T658A,A716G,T1056C,G1171A,A1386G,G1403A,C1408T(4) S
G340A,A579G,G605C,T658A,A716G,T1056C,G1171A,A1386G,G1403A,C1408T(4) S

Clade C: 40,207,605,658,1056,1171,1403,1408,1629
A40G,G207A,G302T,C573A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G515R,G605C,T658A,C816T,T1056C,G1171A,G1396A,G1403A,C1408T,G1629T(4)
A40G,G340A,A579G,G605C,T658A,C672T,T1056C,G1171A,A1386G,G1403A,C1408T,G1624A(4)
A40G,G207A,G605C,T658A,T834A,T1056C,G1171A,G1403A,C1408T,G1473A,G1629T(4)
A40G,G207A,G605C,T658A,G832A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,G1005A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,G1629T(4) F
A40G,G207A,G605C,T658A,T1056C,G1171A,C1182T,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,T834A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,A1435G,G1629T(4) (? F/S)


Odd Ones
C148T,G340A,A424G,A566G,T658A,G684A,C870T,T927C,G1171A,A1230G,G1326A,C1408T,A1536G(4)
T216C,G340A,A424G,G435A,T658A,C672T,G1171A,C1408T(4) F
T243C,G340A,A424G,C473T,T658A,G713A,A747G,G1171A,C1408T,C1627A(4) F
C172T,G360A,A424G,T519A,T658A,G1171A,C1408T(4)
G154A,G207A,G605C,T658A,T1056C,G1171A,C1383A,G1403A,C1408T,G1524K,G1629T(4)
T17C,G565A,T658A,A973G,G1171A,T1197C,G1395A,C1404T,C1408T,G1524A,A1651G(4) F
G396A,G451A,G531A,T598C,T658A,A795G,C921T,A1131G,G1407A,C1408T,C1464T,A1553G,A1579G(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)

By areas
1 Coventry
4 Newcastle-upon-tyne
3 Oxford
1 Durham
2 Leeds
1 Manchester
4 Birmingham
1 Leicester
1 York
1 Cambridge (Fatal)
6 London
2 Nottingham
3 Whitechapel (2 Fatal)
1 Middlesex
1 North Yorkshire

8 Severe/fatal with no locations given
1 Possible severe/fatal no location

 

[NS1]

Re: UK Genetic Evaluation thread

Thank you for evaluating these sequences, mixin.

Bear in mind that a program is most useful when the heuristics and assumptions are fully externalised. We are not in receipt of those information points, so our comments here will be given as one in the dark. We've attempted to gain this required insight in the past without success.

From your narrative and the block of output, we can ascertain that the program did some level of data organisation (unstated) and that you then re-arranged parts of the output based on some secondary heuristics.

Does each line represent one sequence and the nucleotide changes on that sequence? Does the program have a method to effectively label those lines with the sequence names (at the beginning and end)? Does any possibility exist to manage ambiguity codes? Does any possibility exist to notate the amino acid positions in parentheses next to the nucleotide positions [e.g. G451A (137T), T1056C (syn338G), et al] or to provide a separate report notated by amino acid position with silent changes differentiated?

Per line, could we see an automated count of the SNPs (silent and non-syn) and have highlighted the SNPs matching the selected sub-clade standards:

Clade A: 451,598,658,1408,1464
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T[x SNPs, x Silent, x Non-syn] (4)

An output format of .csv would increase ease of visualisation if the changes are properly aligned per dataset. Input into a spreadsheet allows additional automated analyses. Output designed for insertion / update into a relational database is even more useful.

Also note that the sequence deposit does not directly match the Ellis Figure 3 Phylogenetic Tree. So overall count and category counts are very dependent on identifying the individual sequence / sample name.

==

At any rate, the output and the post-processing organisation appears to demonstrate that the selected sub-clades demonstrate a lower CFR than the "Odd Ones"? The odd ones, we suggest, are the sequences that do not pattern onto the major groupings using the unstated parameters and hueristics of the program?

If this statement is true concerning the method of selecting the "Odd Ones", then we may conclude that this category has a CFR of 37.5% (3 of 8) as opposed to much lower death rates on the selected larger groups? In fact, the largest suggested sub-clade has the lowest CFR signalling that transmission v. severity are two different concerns with the current circulating reservoir combinations in the UK.

CFR by Suggested Sub-Clade

37.5% - Odd Ones (3 of 8)
20.0% - Clade B (2 of 10)
20.0% - Clade C (1 or 2 of 10)
08.3% - Clade A (1 of 12)

A higher death rate among unpatterned sequences may potentially signal that the hyper-morphic phase has re-initiated. Our analyses indicate that many of the novel and rare changes incoming to PF11 at this time also exist in animal reservoirs that have been previously related to human infection. Small changes in human influenza matching animal genetics are known to create significant changes in severity levels and viral behaviour.

Hyper-morphism (high change rate) with a significant percentage of animal-matched intake equates to a present hyper-zoonotic set of strains.

This hypothesis appears to agree with your data . . . if we are guessing correctly at the underlying parameters.

 

[gsgs]

Re: UK Genetic Evaluation thread

I just found this thread.
When will the sequences be published so we can
verify and discuss it freely ?

Earlier attempts to assign differences in virulence to
different substrains had failed.
Remember Argentina last year, Ukraine, early Mexico.

A complete statistic with the proportion of deaths in the
published (or GISAID) substrains would be useful, though.

calculating % of deaths (or ICU) per mutation is straight forward,
then assign the sums of the values to each virus
then calculate the average for each month and chart
it over time, per country,continent

it has been speculated that more middle-age people are infected
this season vs. more children last season.
And the children die less often from ***, more are immune now,
and that that allone could explain the increased number of deaths.

Do we have UK-deaths by age 2009/10 vs. 2010/11 ?

Virologically it's remarkable that these substrains are new,
they emerged from early Cancun viruses, presumably in Asia
and have nothing in common with the strains circulating
in USA,Europe last season.

These mutations do not occur e.g. in my list of 28 substrains
from Nov.2010

 

[NS1]

Re: UK Genetic Evaluation thread

I just found this thread.
When will the sequences be published so we can
verify and discuss it freely ?

According to the narrative, these sequences have apparently been
processed through a program of your authorship?

Earlier attempts to assign differences in virulence to
different substrains had failed.
Remember Argentina last year, Ukraine, early Mexico.

Clear indications of severity markers and Vaccine Escape domains are available as are indications of drug resistance.

A complete statistic with the proportion of deaths in the
published (or GISAID) substrains would be useful, though.

Some outdated material has been produced and analysed to little effect due to scarcity and unreliable parameters.

calculating % of deaths (or ICU) per mutation is straight forward,
then assign the sums of the values to each virus
then calculate the average for each month and chart
it over time, per country,continent

Such small point factors are unreliable for many reasons, but, most importantly, due to various synergies among combinations of polymorphisms, some that produce little net effect and others that produce multiplicative effect. The record is observable on these matters.

Calculations based on selected combinations of polymorphisms chosen by frequency / absence on a background among other factors may provide comparisons that feedback to the choice of SNPs to include on that calculation. Due to sparsity of data and clinical meta-data, the calculations must be iterated and interpreted with a high level of earnestness and objectivity.

it has been speculated that more middle-age people are infected
this season vs. more children last season.
And the children die less often from ***, more are immune now,
and that that allone could explain the increased number of deaths.

Do we have UK-deaths by age 2009/10 vs. 2010/11 ?

The data have been produced by HPA and provided in spreadsheet format. Transparency and timely reporting is not occurring, nor do we expect valid data in the near term.

GeneWurx_UK_ICU_Categories_v1.xls

Virologically it's remarkable that these substrains are new,
they emerged from early Cancun viruses, presumably in Asia
and have nothing in common with the strains circulating
in USA,Europe last season.

These mutations do not occur e.g. in my list of 28 substrains
from Nov.2010

Though novelty and divergency is now the norm in the UK reservoir, one of the strains that was previously found in a robust form during the late Australia winter season is easily trackable.

Emergence began in late 2009 in the United States of the most frequent background and continued via domaining during the early part of 2010 and the summer in the Northern Hemisphere, translating well into the winter in the Southern Hemisphere.

GeneWurx analyses show a pattern of emergence throughout the world of this particular currently low CFR strain. What will happen next may be more predictable should sequences be provided in a timely fashion in conjunction with mild, severe and fatal cases.

GeneWurx_UK_December_Emerging_Genetics_v3.xls

 

[gsgs]

Re: UK Genetic Evaluation thread

> According to the narrative, these sequences have apparently been
> processed through a program of your authorship?

mixin uses my mutatm.exe to extract the mutations
several other programs are at http://magictour.free.fr/seq
or on request

> Clear indications of severity markers and Vaccine Escape domains are available as are
> indications of drug resistance.

these are not specific to substrains but occur independently on several backgrounds

> Some outdated material has been produced and analysed to little effect due to scarcity
> and unreliable parameters.

sequencing gets plentiful and cheap ... someone should do a better analysis and publish it

> Such small point factors are unreliable for many reasons, but, most importantly, due to
> various synergies among combinations of polymorphisms, some that produce little net
> effect and others that produce multiplicative effect. The record is observable on these matters.

we do what we can. I have not yet seen that analysis.
Once the data is available we can also easily test it for combinations

> Calculations based on selected combinations of polymorphisms chosen by frequency /
> absence on a background among other factors may provide comparisons that feedback
> to the choice of SNPs to include on that calculation. Due to sparsity of data and clinical
> meta-data, the calculations must be iterated and interpreted with a high level of
> earnestness and objectivity.

should be possible. We should also include info on the antigenic situation of
the population before infection

> The data have been produced by HPA and provided in spreadsheet format.
> Transparency and timely reporting is not occurring, nor do we expect valid
> data in the near term.

average age this season = ?
average age last season = ?

> Though novelty and divergency is now the norm in the UK reservoir, one of the strains
> that was previously found in a robust form during the late Australia winter season is
> easily trackable.

which one ? (use mixin letters above)

> Emergence began in late 2009 in the United States of the most frequent background and
> continued via domaining during the early part of 2010 and the summer in the
> Northern Hemisphere, translating well into the winter in the Southern Hemisphere.

did it "emerge" and spread from USA in late 2009 ? Where was it before ?

 

[gsgs]

Re: UK Genetic Evaluation thread

Clade A: 451,598,1464
Clade B: 340,1056,1171,1403
Clade C: 40,207,605,1056,1171,1403,1629


mutations with frequency occurring in my list of 28 substrains from Nov.2009
(my enumeration starts with the coding region, so divide by 3 to get the amino acid position)

G0298A(5),12
G0316A(6),12
G1143A(5),12
A0742G(6),10
C1408T(4),10
A1741C(3),7
C1118T(5),7
A0367G(8),6
G0492A(7),6
G0600A(7),6
G1248A(5),6
G2163A(1),6
T0658A(4),6
G0546A(1),5
C0145A(4),4
A0300G(2),3
A1058G(2),3
A0283G(6),2
A1218G(1),2
A1577G(1),2
C0696T(7),2
C0954T(6),2
C1332T(2),2
G0037A(4),2
G0892A(7),2
G1173A(4),2
G1986T(3),2
A0004G(4),1
A0008G(3),1
A0177T(1),1
A0189G(8),1
A0208G(8),1
A0273G(1),1
A0366G(2),1
A0456G(3),1
A0541G(2),1
A0561C(3),1
A0588G(3),1
A0716G(4),1
A0723G(4),1
A0775G(8),1
A0877G(4),1
A0898G(4),1
A0923G(6),1
A1044G(6),1
A1051G(1),1
A1381G(1),1
A1429G(4),1
A1467T(2),1
A1603G(2),1
A1674G(1),1
A1845G(3),1
A1920G(2),1
A1979G(1),1
A2030G(1),1
C0022A(4),1
C0132T(5),1
C0148T(4),1
C0163A(8),1
C0259T(4),1
C0334T(1),1
C0397A(5),1
C0480T(3),1
C0547A(1),1
C0643T(8),1
C0717A(7),1
C0738A(8),1
C0870T(4),1
C0888T(1),1
C0896T(4),1
C1140T(4),1
C1263T(3),1
C1389A(5),1
C1434T(5),1
C1945T(3),1
C2076T(2),1
G0110C(1),1
G0165A(7),1
G0201A(5),1
G0268T(4),1
G0279A(8),1
G0282A(1),1
G0477A(3),1
G0480A(1),1
G0522A(7),1
G0571A(2),1
G0576A(4),1
G0603A(2),1
G0624A(5),1
G0636A(2),1
G0659A(8),1
G0687T(4),1
G0705A(2),1
G0724A(8),1
G0786A(6),1
G0797A(3),1
G0799A(1),1
G0891A(4),1
G0930T(4),1
G0936A(3),1
G1012A(4),1
G1097A(6),1
G1163A(3),1
G1308A(1),1
G1438A(1),1
G1563A(4),1
G1680A(2),1
G1945A(1),1
G1968A(3),1
G2019A(3),1
G2091A(2),1
T0267C(5),1
T0298C(4),1
T0435A(3),1
T0450C(2),1
T0480C(5),1
T0670C(3),1
T0711C(6),1
T0717C(4),1
T0774C(1),1
T0933C(4),1
T1150C(2),1
T1275C(2),1
T1305C(3),1
T1335C(3),1
T1354C(4),1
T1527C(3),1
T1626C(1),1
T1711C(1),1
T1760G(2),1
T1854C(1),1
T1857C(2),1
T1872C(1),1
T1899C(1),1
T2000C(2),1
T2189C(2),1
T2241C(2),1
A1044G(6),1

 

[NS1]

UK Severe Wave Sequences Show No Fatalities with 225G

UK Severe Wave Sequences Show No Fatalities with 225G

UK Severe Wave Sequences Show No Fatalities with 225G

The UK Health Protection Agency released a group of 37 sequences at GISAID dated 2010-01-05 that appeared sometime after that date. Some of the sequences in this deposit also appear on the Figure 3 Phylogenetic Tree from the HPA Eurosurveilance paper requested to be cited as:

Ellis J, Galiano M, Pebody R, Lackenby A, Thompson C, Bermingham A, McLean E, Zhao H, Bolotin S, Dar O, Watson JM, Zambon M. Virological analysis of fatal influenza cases in the United Kingdom during the early wave of influenza in winter 2010/11. Euro Surveill. 2011;16(1)ii=19760.
Available online:
http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19760
Date of submission: 31 December 2010

The science community expected a full deposit to back the paper, but not all sequences noted on the phylogenetic table were in place at GISAID as of this 2nd deposit.

The recent GISAID deposit added 4 of the fatalities noted on the Ellis Figure3 <SUP>1</SUP> phylogenetic tree to the 4 fatalities previously released on 2010-12-20. The previously released sequences were not marked at GISAID with a fatality notation, but now have been notated as "Deceased" as of yesterday. Oddly enough, on the 2010-01-05 deposit, none of the sequences notated as fatal on the phylogenetic tree were marked as to outcome at GISAID? This dys-synchronicity and cycling of outcome data is less than optimal for worldwide communication.

None of the sequences at the GISAID database that correspond to notated fatalities on the Eurosurveillance paper carry the severity marker of HA 225G.

225G is found on identical HA sequences noted as 2 severe cases.

• UKEngland4940476_2010_12
  
• UKEngland4880378_2010_12
  

. . . . UKEngland4940476_2010_12 (
. . . . . . . . 100N,
. . . . . . . . syn179L (CTg) [Regional Marker 2009 (tTA)]
. . . . . . . . . . . . . . . . . . . . [TexasAF2588_2009_10_04,
. . . . . . . . . . . . . . . . . . . . TexasJMS404_2010_01_08],
. . . . . . . . 188T,
. . . . . . . . 225G,
. . . . . . . . syn338G,
. . . . . . . . 377K,
. . . . . . . . syn448L [Regional Marker 2009]
. . . . . . . . . . . . . . . [UKEngland4880378_2010_12 with syn179L, 188T,
. . . . . . . . . . . . . . . UKEngland4640543_2010_11_f with syn179L, 188T, 190G
. . . . . . . . . . . . . . . UKEngland4920303_2010_11 with syn179L, 188T,
. . . . . . . . . . . . . . . UKEngland142_2010_11 with syn179L, 188T,
. . . . . . . . . . . . . . . UKEngland4860049_2010_11 with syn179L, 188T,
. . . . . . . . . . . . . . . UKEngland126_2010_11 with syn179L, 188T,
. . . . . . . . . . . . . . . NZChristchurch8_2010_07_08
. . . . . . . . . . . . . . . . . . . with syn44L, 97N, syn99I, syn106E, 128D,
. . . . . . . . . . . . . . . . . . . . . . . syn214K, 253A, syn362S, 377K,
. . . . . . . . . . . . . . . Calif06_2010_04_05 with 269V
. . . . . . . . . . . . . . . Hawaii08_2010_04_12
. . . . . . . . . . . . . . . . . . . with 159D, 269V, 312R, 313K,
. . . . . . . . . . . . . . . Ethiopia13_2010_02_10
. . . . . . . . . . . . . . . . . . . with 100N, syn163K, 269V, 324I, syn360Q, syn455Q,
. . . . . . . . . . . . . . . Philippines824_2010_02_17 with 165N,
. . . . . . . . . . . . . . . Kosova876_2009_12_22,
. . . . . . . . . . . . . . . RussiaYakutsk_EAV_2009_11_18,
. . . . . . . . . . . . . . . Netherlands2143_2009_11_16 with syn179L (tTA),
. . . . . . . . . . . . . . . RussiaYaroslavl_CHMV_2009_11_10_f with 224K & 225G mix
. . . . . . . . . . . . . . . AntwerpINS221_2009_10_28 with syn179L (tTA)],
. . . . . . . . 454N)

. . . . UKEngland4880378_2010_12 (
. . . . . . . . 100N,
. . . . . . . . syn179L,
. . . . . . . . 188T,
. . . . . . . . 225G,
. . . . . . . . syn338G,
. . . . . . . . 377K,
. . . . . . . . syn448L,
. . . . . . . . 454N)

With 8 fatalities noted on the tree and none of them coming from cases manifesting an HA 225G severity marker, we must ask for more data while looking elsewhere for causality. The common theme in recent UK and other geographies is hyper-zoonosis creating divergency. Animal reservoirs (H3N8, H5N1, H7N7) carry change values matching the recent intake into the pH1N1 circulation. Data suggests that this human disease is rapidly diversifying by accumulating genetic information from multiple animal reservoirs.

For example, the 377G that is emergent in 2009 Human H5N1 from Dakahlia, Egypt is also emergent in England and Iran in the Pandemic H1N1 reservoir. A recent fatal case from England, UKEngland4500186_2010_11_f, carries the same coding for 377G combined with a 190Y that is found elsewhere in 2010 Malmoe, Sweden pH1N1 and in Avian H6N1.

The HPA Ellis Figure3 <SUP>1</SUP> phylogenetic tree did not notate the 377G branch though 5 English sequences, including this fatal one, appear on the chart carrying the H5N1 Human adaptation.

• UKEngland83_2010_10
  
• UKEngland87_2010_10
  
• UKEngland5500192_2010_10
  
• UKEngland119_2010_11
  
• UKEngland4500186_2010_11_f
  

GeneWurx has annotated the Ellis Figure3 <SUP>1</SUP> phylogenetic tree with green boxes next to the four fatal GISAID 2010-12-20 deposited sequences and has proposed polymorphism notations on various unmarked branches in an attempt to provide clarity.

UK_2010_Phylo_ELLIS_Fig3new_4_GISAID_Notated_2011_01_08.JPG

Is the Hydra Effect providing Immune Escape capacity for this circulating viral reservoir? Has an accelerant been insinuated into the reservoir?

In many locales around the United Kingdom, half of the ICU beds are occupied by Influenza patients. Only 1 of 12 ICU beds was occupied by an Influenza patient at the peak of last year's pandemic. 100% of the ECMO capacity (heart/lung machine) has been utilised for more than 2 weeks. Hospitals have moved to "Black Alert" status. Ireland has set new records for the number of ER (Casualty, A&E) patients left on gurneys due to room capacity problems.

Do these facts move the HPA to quickly derive plausible causality and workable next steps?

Usable data streams with matched clinicals to the sequences are the primary inputs required to investigate and validate causality. An explanation is deserved from the HPA on their removal of this valuable data from the science community?


1. Ellis J, Galiano M, Pebody R, Lackenby A, Thompson C, Bermingham A, McLean E, Zhao H, Bolotin S, Dar O, Watson JM, Zambon M. Virological analysis of fatal influenza cases in the United Kingdom during the early wave of influenza in winter 2010/11. Euro Surveill. 2011;16(1)ii=19760. Available online: http://www.eurosurveillance.org/ViewArticle.aspx?ArticleId=19760

 

[mixin]

Re: UK Genetic Evaluation thread

To clarify: each line represents the nucleotide changes in a given sequence; I didn't manipulate the individual changes. An example is England/87. The (4) at the end just verifies that I'm doing segment 4 (HA)

A/England/87/2010
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)

I like to look at the nucleotides so I can see all the changes and not just the ones on the amino acid level. I omitted the name of each sequence for space purposes; I wanted to compare just the changes to see if I could identify the emerging subclades.

After I did all the sequences in the above format, I then decided to group them by areas and I saw a bit of a trend. Here's an example:

Newcastle-upon-tyne 3 in Oct, 1 Nov
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)

Oxford all in Oct
G451A,T598C,T658A,T858C,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)

But, beyond those 2 areas, the clades seemed quite diverse. See here with London:

London 1 Aug, 1 Nov, 4 Dec
C172T,G360A,A424G,T519A,T658A,G1171A,C1408T(4)
C148T,G340A,A424G,A566G,T658A,G684A,C870T,T927C,G1171A,A1230G,G1326A,C1408T,A1536G(4)
A40G,G207A,G302T,C573A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
G154A,G207A,G605C,T658A,T1056C,G1171A,C1383A,G1403A,C1408T,G1524K,G1629T(4)
A40G,G207A,G605C,T658A,G832A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,G1005A,T1056C,G1171A,G1403A,C1408T,G1629T(4)

I noticed that many of the sequences had a first mutation in common (G451A in Newcastle-upon-tyne); so I just started grouping them together.

I'll be glad to post everything I have regarding the UK sequences; but when I work with GISAID sequences, I'm never absolutely sure how much info I post is acceptable. My understanding is that as long as I don't post the actual sequence, I'm within their rules to post the mutations and other information on the submission page.

 

[NS1]

Divergence

Divergence

To clarify: each line represents the nucleotide changes in a given sequence; I didn't manipulate the individual changes. An example is England/87. The (4) at the end just verifies that I'm doing segment 4 (HA)

A/England/87/2010
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)

I like to look at the nucleotides so I can see all the changes and not just the ones on the amino acid level. I omitted the name of each sequence for space purposes; I wanted to compare just the changes to see if I could identify the emerging subclades.

After I did all the sequences in the above format, I then decided to group them by areas and I saw a bit of a trend. Here's an example:

Newcastle-upon-tyne 3 in Oct, 1 Nov
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,C1408T,C1464T(4)
G451A,T598C,T658A,T933C,T1020C,T1191C,G1206A,C1245T,T1374G,G1403A,C1408T,C1464T(4)

Oxford all in Oct
G451A,T598C,T658A,T858C,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)
G451A,T598C,T658A,T879C,C1098T,A1172G,G1266A,C1408T,C1464T,A1568G,G1577T,G1630A(4)

But, beyond those 2 areas, the clades seemed quite diverse. See here with London:

London 1 Aug, 1 Nov, 4 Dec
C172T,G360A,A424G,T519A,T658A,G1171A,C1408T(4)
C148T,G340A,A424G,A566G,T658A,G684A,C870T,T927C,G1171A,A1230G,G1326A,C1408T,A1536G(4)
A40G,G207A,G302T,C573A,G605C,T658A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
G154A,G207A,G605C,T658A,T1056C,G1171A,C1383A,G1403A,C1408T,G1524K,G1629T(4)
A40G,G207A,G605C,T658A,G832A,T1056C,G1171A,G1403A,C1408T,G1629T(4)
A40G,G207A,G605C,T658A,G1005A,T1056C,G1171A,G1403A,C1408T,G1629T(4)

I noticed that many of the sequences had a first mutation in common (G451A in Newcastle-upon-tyne); so I just started grouping them together.

I'll be glad to post everything I have regarding the UK sequences; but when I work with GISAID sequences, I'm never absolutely sure how much info I post is acceptable. My understanding is that as long as I don't post the actual sequence, I'm within their rules to post the mutations and other information on the submission page.

We've also found those two combinations in tandem around the world. The novel and rare inputs to the others compel us, however, to look beyond the points that coalesce to the points of divergence.

The GeneWurx method also requires examination of the nucleotide level and places emphasis on the silent changes due to various unexplained behaviours we've noted in the sequence record. The results of the nucleotide examination are then normed to the amino acid level for reporting by preceding any synonymous changes with "syn", e.g. syn254P. Most papers are already difficult enough to match due to various numbering schemes, so using the amino acid level reporting allows a higher match ratio to previously published research (sparse database).

We'd enjoy having an additional stream of data analysis, but do not have the resources at this time to continually re-interpret the numbering, labeling and heuristics of this particular program.

 

[mixin]

Re: UK Genetic Evaluation thread

We'd enjoy having an additional stream of data analysis, but do not have the resources at this time to continually re-interpret the numbering, labeling and heuristics of this particular program.

That's ok. I really didn't expect re-interpretation of all the numbering, labeling and heuristics. I assumed that you were also looking at the nucleotides and came up with similar results.

Thanks for looking.

 

[NS1]

Re: UK Genetic Evaluation thread

That's ok. I really didn't expect re-interpretation of all the numbering, labeling and heuristics. I assumed that you were also looking at the nucleotides and came up with similar results.

Thanks for looking.

We were lamenting that the program heuristics are not externalised and that the numbering is not produced in both formats for usability and for comparison. Having more eyes on the object generally provides a better description if the language is agreed and externalised. Providing an evaluation when the data structures, the data labels and the manipulation heuristics are not externalised would be of questionable success, so we asked about the potential.

The data is being addressed at the nucleotide level, an importance that we emphasised when discovering the cross-linkage of the first Vaccine Escape sequence (UkraineLvivN6 with HA 225G, HA syn413K and NA syn407V).

You're doing a great job of digging into this data with the tools available, mixin! Please keep up the good work.

 

[JJackson]

Re: UK Genetic Evaluation thread

I have posted some thoughts on the current situation here (not wanting to clog this news thread with too much speculation on the meaning of the available sequence data). It in part addresses Toaster2’s question and addresses some points raised in posts by ironorehopper and NS1.

The following is a personal opinion only and as I am neither a virologist nor MD, caveat emptor.

On the broad question ‘do the recorded SNPs say anything interesting about virulence or epidemiology more generally?’ I would say no, at least not yet. This is not due to a lack of effort on the part of NS1, or anyone else, but there is just not enough timely data. To be able to draw useful conclusions we need sequence data – not just partial, and not just HA & NA – with linked anonamised (if that is a word) clinical data and we need it in quantity and in as near real-time as possible. This does not exist partly because it is not collected in this form – linked to clinical records - and partly because the sequence data is often used by the researchers, for their own research, before release months or even years later. The whole problem is compounded by the fact that only the currently active tips of the relevant phylogenic trees are of any interest.

The really useful data is that collected from the sentinel GP practises. Sequences are primarily generated by two routes. If you catch flu this happens
1] You do not go to the doctor but stay home and self medicate – numerically the biggest group.
2] You go to the GP, this is now a self selecting group and will be skewed towards the more clinically severe cases. You may be prescribed anti-virals.
2a] Your GP is not in the sentinel surgery group – vast majority – and you are not tested.
2b] Your GP is in the flu surveillance group and a sample is collected.
3] Regardless of whether you were in 1] or 2] you may not get better and may present to hospital or go back to option 2] (see your GP) and have a sample taken and/or be given antivirals.
4] While in hospital you may have multiple samples taken and analysed.
All of the above would be much better as a flowchart but I haven’t the software.

There are two observation I would like to make regarding the system.
Firstly regarding Antivirals - apart from late stage step 4] - this means Tamiflu in nearly all cases. Tamiflu needs to be given ‘within 48hrs of symptom onset’ - as per the pack -the only group who could realistically meet this criteria are the subset of group 2] who were prescribed Tamiflu at their first GP visit. (this is the argument behind my earlier reply to mixin’s surprise at the often late use of Tamiflu)
Secondly regarding samples. Unfortunately the best (most random and therefore most indicative of the circulating strain) samples come from group 1] which never get tested. 2b] is the next best group in that they are fairly random but skewed to excluded asymptomatic and mildly symptomatic cases.
Once we get into groups 3] & 4] we start seriously skewing the data.
They are disproportionately likely to have
a] a sample taken
b] more than one sample taken
c] a sample taken after they have antivirals in their system

Route c] will apply selective pressure favouring any resistant strains in the host’s quasi-species mix and a] & b] will skew the total number of samples towards the hospitalised patients. Finally the likelihood of key clinical data being matched to a given sample is increased if the case is fatal or clinically note worthy. This also increases the chance of a given swab being sequenced as swabs may be taken but not bumped up the sequencing cue unless the patient is failing to respond to treatment (clinician wanting to rule out antiviral resistance development, again a route to multiple sequences from a single host).

I hope I have shown why it would be useful to be able to separate the 2b] data from the 3] & 4] data. The route 3] & 4] data is from the severe cases and may be different from the 2b] data. If it can be shown to be genetically different the cause could be either because one, or more, of the circulating strains are more virulent or because in those patients who fail to clear the disease unaided, and end up in hospital, the predominant stain in the patient changes over time. To sort out these scenarios we would need a series of sequences from 2b] patients who later ended up in ICU so the genetic mix could be observed over the course of the illness. There have been data series on immunocompromised showing resistance development but I have not seen any on ‘ordinary’ patients who were tested with mild illness and followed the sequences through to severe illness or death, and we would need several, preferably all from the same genetic starting point, for significance (in its mathematical sense).

In post No.302 of the main UK news thread ironorehopper posted a couple of phylogenic trees from HPA analysis of English data mainly from 2010. The first thing to notice – in line with the argument made above – is the proportion of severe and fatal cases (see footnotes for key) as a proportion of the included sequences. Obviously this is not representative of the average virus causing community illness or the Case Fatality Rate would be 1918 like, or worse.
The next thing I would draw your attention to is the very bottom sample (Califronia/7) on which the vaccine is based. I would have preferred to see this nestled more within the main branch structure with shorter average horizontal distances to the other sequences. (I have written a little about phylogenic trees in a box at the end of the post for those that are unfamiliar with them). I have not seen any compelling evidence for vaccine escape, although I would probably advocate changing from California/7 to something selected from the centre of the most active branch for subsequent seasons as Cal/7 seems to be from an extinct branch – although still close enough to be immunologically useful so far.

Re Vaccine Escape
In Nov 09 I posted my scepticism (Post number 21 of what became the ‘Low Reactor’ thread) regarding vaccine escape due to a SNP – or even several SNPs – which I viewed, and still view, as impossible. There is one, very unlikely, scenario in which that statement might not be true - that I can think of – and as no one raised it in the ensuing debate (not everyone agreed with me – hard thought that is to believe) I covered it in post 157, with some further explanation in post 156. As this ended up being an enormous, and at times heated, thread (I alone ended up making 25 posts defending my position) I will not go over all the arguments again as my position has not changed.

Re antiviral resistance
This is taken from the same eurosurveillance paper the phylogenic trees came from

Between October and December 2010, antiviral resistance monitoring was undertaken on 156 community and 159 hospital isolates. Six cases of oseltamivir resistance associated with the H275Y mutation in the neuraminidase (NA) gene have been detected. Only one of these cases has had known exposure to oseltamivir, Two of them have been identified from community surveillance of uncomplicated infections, three cases have been detected before treatment in individuals hospitalised with underlying risk factors, and the sixth case has been detected after oseltamivir treatment in a hospitalised individual.

The Italics are mine.
While their have always been some resistant sequences the shift from immunocompromised patients to the current situation with 2 out of 156 resistant - from the group I earlier described as 2b] – is much more worrying. Once H275Y found a wild type background on which it suffered no significant fitness penalty (in the now largely extinct seasonal H1N1) it supplanted the susceptible wild type in fairly short order despite no great Tamiflu usage in the areas where it proliferated. This does not bode well for our only realistic antiviral. Yes I know there are others but inhaled powders are not ideal for a respiratory illness and IV is fine in a hospital setting but not very practical elsewhere and none are available in the same volumes if Tamiflu is lost for H1N1(2009).

As promised some basics on Phylogenic trees (AKA Cladograms)
Firstly do not worry that Cal/7 is at the bottom edge of the table as the branches can be articulated about their nodes for convenience of display and labelling – vertical position is not relevant.
Each sequence is compared to every other sequence and as more differences are found their horizontal spacing is increased. For our purposes do not need to worry about the actual lengths of the lines (some of which have a length on them) to get a ‘feel’ for the relative closeness, or remoteness, of sequences look at the comparative lengths.
As with ordinary xy graphs the scale can be adjusted for conveyance to fit the range for any given data set. For instance if you are plotting only H1N1(2009) sequences from England in 2010 they are all going to be relatively close to one another, if we added a seasonal H1N1 from 2006 and H1N1(1918) then all the H1N1(2009)s would form a dense clump on one branch with the other two having their own branches. The net result would be a 3 pronged fork with one sequence each on prongs one and two and clump of sequences on third (H1N1(2009)) prong. Also there are different algorithims for generating the lengths in different cladogram types (analogous to using a log rather than linier scale on one axis).

The problem with the tree – like the ones in the HPA paper – is they plot only distance not time (beyond most having 2010 in their name) what we need (nice little project for any of you software coders) is to add a temporal dimension. This could be done as an animation so you could ‘see’ the tree grow over time with the individual sequences ‘flowering’ on their bare branches as time runs forward at a second/month scale. Alternatively colour could be use with the earliest sequences printed in blue and running through the rainbow with the last sequence in red. The problem this would solve is that when looking at a tree we may see a prominent branch, with many sequences, but it may in fact represent an evolutionary dead end. Had we had our rainbow colours we might have seen the whole branch as shades of blue (implying it had died out) while an apparently insignificant branch at the other end of the cladogram, which was predominantly red, would be the place to pick our vaccine target for next year.

 

[gsgs]

Re: UK Genetic Evaluation thread

could I, if I thought I had flu, collect a sample by myself, send it to a lab
and get the sequences and publish them ?

What would it cost ?
Or would that be illegal in UK ?

 

[mixin]

Re: UK Genetic Evaluation thread

JJackson,
Speaking of low reactors, I'm curious as to your opinion of a statement in the last MRC report.

If I'm understanding this correctly, the problem with the Victoria-lineage is not so much with the vax virus but more with the medium it was grown in? If the vax strain was grown in mammal cells, then it would have a strong reaction?

In the Victoria-lineage many viruses yielded reduced titres with anti-sera raised against the egg-grown vaccine virus. Anti-sera raised against reference strains isolated and propagated in mammalian cells continued to react well with recently isolated viruses.