J. Biomedical Science and Engineering, 2010, 3, 584-601
doi:10.4236/jbise.2010.36081 Published Online June 2010 (http://www.SciRP.org/journal/jbise/ JBiSE ).
Published Online June 2010 in SciRes.http://www.scirp.org/journal/jbise
Novel host markers in the 2009 pandemic H1N1 influenza a
virus
Wei Hu
Department of Computer Science Houghton College, Houghton, USA.
Received 21 March 2010; revised 28 April 2010; accepted 8 May 2010.
ABSTRACT
The winter of 2009 witnessed the concurrent spread
of 2009 pandemic H1N1 with 2009 seasonal H1N1. It
is clinically important to develop knowledge of the
key features of these two different viruses that make
them unique. A robust pattern recognition technique,
Random Forests, was employed to uncover essential
amino acid markers to differentiate the two viruses.
Some of these markers were also part of the previ-
ously discovered genomic signature that separate
avian or swine from human viruses. Much research
to date in search of host markers in 2009 pandemic
H1N1 has been primarily limited in the context of
traditional markers of avian-human or swine-human
host shifts. However, many of the molecular markers
for adaptation to human hosts or to the emergence of
a pandemic virus do not exist in 2009 pandemic
H1N1, implying that other previously unrecognized
molecular determinants are accountable for its capa-
bility to infect humans. The current study aimed to
explore novel host markers in the proteins of 2009
pandemic H1N1 that were not present in those clas-
sical markers, thus providing fresh and unique in-
sight into the adaptive genetic modifications that
could lead to the generation of this new virus. Ran-
dom Forests were used to find 18 such markers in
HA, 15 in NA, 9 in PB2, 11 in PB1, 13 in PA, 10 in
NS1, 1 in NS2, 11 in NP, 3 in M1, and 1 in M2. The
amino acids at many of these novel sites in 2009 pan-
demic H1N1 were distinct from those in avian, hu-
man, and swine viruses that were identical at these
positions, reflecting the uniqueness of these novel
sites.
Keywords: 2009 Pandemic H1N1; Host Switch; Influ-
enza; Mutation; Random Forests
1. INTRODUCTION
In addition to the common seasonal H1N1 influenza
virus, an antigenically novel swine-origin pandemic
H1N1 influenza virus marked the flu season in 2009. It
is likely that both 2009 pandemic H1N1 and seasonal
influenza will coexist for some time. Elucidation of the
characteristics of this new virus has become an impor-
tant part of the current flu research. The identification of
molecular markers for drug resistance, virulence, viral
transmission and replication, human adaptation, and
evolution can shed new light into the nature of this virus.
There are eight single-stranded RNA segments of the
influenza A virus genome. They code 11 proteins: hema-
gglutinin (HA), neuraminidase (NA), matrix 1 (M1),
matrix 2 (M2), nucleoprotein (NP), non-structural pro-
tein 1 (NS1), non-structural protein 2 (NS2; also termed
nuclear export protein, NEP), polymerase acidic protein
(PA), polymerase basic protein 1 (PB1), polymerase
basic protein 2 (PB2), and polymerase basic protein 1 –
F2 (PB1-F2). Segments 1, 3, 4, 5, and 6 each encode a
single protein, i.e., PB2, PA, HA, NP, and NA, respec-
tively, whereas segments 2, 7, and 8 each encode two
proteins, i.e. , PB1 and PB1-F2, M1 and M2, NS1 and
NS2, respectively. The life cycle of influenza virus has
the following steps with several proteins involved in
each: entry into the host cell (HA, M1 and M2), entry of
viral ribonucleoproteins (vRNP) into the nucleus (NP,
PA, PB1 and PB2), transcription and replication of the
viral genome (PA, PB1, PB2, NS1, and NP), export of
the vRNPs from the nucleus (NP, NS2 and M1), and
assembly and budding at the host cell plasma membrane
(HA, NA, M 1 a nd M2) [1].
Besides mutations, viruses with segmented genomes
can generate genetic diversity by exchanging gene seg-
ments between different viruses to produce a new virus.
Comprehensive phylogenetic analysis suggested that the
genes of 2009 pandemic H1N1 were derived from avian
(PB2 and PA), human H3N2 (PB1), classical swine (HA,
NP and NS), and Eurasian avian-like swine H1N1 (NA
and M) lineages [2].
The symptoms of the 2009 pandemic H1N1 flu are

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 585
similar to the 2009 seasonal flu with a possibility of ad-
ditional symptoms such as vomiting and diarrhea [3].
The Center for Disease Control and Prevention (CDC)
and Mayo Clinic have developed several molecular tests
to detect and discriminate the novel 2009 pandemic
H1N1 virus and the 2009 seasonal virus [4]. The matrix
(M) gene is highly conserved compared to other gene
segments, which makes it an ideal target for RT-PCR
assays used to detect the presence of influenza. A muta-
tion in the M gene of the 2009 pandemic virus could
invalidate these tests [5].
Sequence survey suggested that there were two dis-
tinct evolutionary trends in antig enic drift of H1N1 HAs
at two residues 190 and 225. The epidemic H1N1 HAs
favor position 190 while the 1918 pandemic and swine
HAs favor position 225 [6]. In contrast to these two
trends, the 2009 pandemic H1N1 strains are highly con-
served at both HA 190 an d 225 and possess th e signature
markers Asp190 and Asp225 that are known to confer
specificity to the human α2-6 sialylated glycan receptors
[7]. Further analysis indicated the 2009 pandemic H1N1
HAs possess residues that can be positioned to bind to
avian α2-3 sialylated glycan receptors as well [7,8]. By
homology modeling of the HA structure, the antigenic
similarity between the 1918 H1N1 and the pandemic
2009 H1N1 viruses was confirmed, and the future amino
acid substitutions on the antigenic sites of 2009 pan-
demic H1N1 HA were also predicted [9], raising the
concerns that these two pandemic H1N1 viruses may
share a similar evolutionary path. With informational
spectrum method [10], a bioinformatics technique, h ighly
conserved domains and mutations in the 2009 pandemic
H1N1 HAs were identified and the contributions of these
mutations to the changes of binding specificity of the
2009 pandem i c H1N1 HAs w e re quantified [11- 13] .
Many 2009 seasonal H1N1 strains carry a NA muta-
tion H275Y that confers high-level resistance to osel-
tamivir. Although most 2009 p andemic H1N1 strains are
susceptible to oseltamivir, the co-circulation of pandemic
and seasonal H1N1 viruses might provide opportunities
for 2009 pandemic H1N1 to develop oseltamivir resis-
tance through mutations and reassortments between
pandemic and seasonal H1N1 viruses. Positive natural
selection was detected in the NA proteins of 2009 pan-
demic H1N1 at codons 275 and 248 and seasonal H1N1
at codon 275, with statistically significant bias of non-
synonymous mutations relative to synonymous muta-
tions [14]. Besides positio n 275, mutations at other posi-
tions in NA such as 116, 117, 119, 136, 150, 151, 199,
223, 275, and 295 could also alter NA inhibitor suscep-
tibility [15,16].
Two recent reports [17,18] revealed three NA variant
groups in 2009 pandemic H1N1. The first group had
V106 and N248, the second included I106 and N248,
and the third contained I106 and D248, highlighting the
rapid genetic variation of this surface antigen under host
immune pressure and the need for close monitoring. The
NA protein of the avian viruses has, in addition to the
catalytic site, a separate sialic acid binding site that is
not present in human viruses, which could enhance the
catalytic efficiency of NA [19]. Although the second
binding site was not conserved in swine NA strains, a
recent report found the 2009 pandemic H1N1 strain of
swine origin appeared to have retained some of the key
features of the second binding site. Their data showed
possible lowered HA activity for this second site, which
might be an important event in the emergence of the
2009 pandemi c strain [20] .
The interaction of NP and the influenza polymerase,
containing the PA, PB1 and PB2 proteins, catalyses viral
RNA replication (vRNA→cRNA→vRNA) and transcrip-
tion (vRNA→mRNA) in the nucleus of infected cells.
The PB2 protein of human viruses tend to possess a ly-
sine at position 627 (K627), whereas avian viruses gen-
erally have glutamic acid at this position (E627). The
mutation E627K allows avian virus to efficiently grow in
humans and was identified experimentally as a crucial
host range and pathogenicity determinant [21,22]. The
2009 pandemic H1N1 strains could transmit in humans
efficiently, but exclusively possess the avian signature
E627. Therefore, there might be alternative strategies
employed by the novel 2009 H1N1 polymerase to main-
tain the efficient replication rate. A recent study discov-
ered that serine at position 590 (S590) and arginine at
position 591 (R591) might serve as a regulator of poly-
merase activity that contributes to the increased replica-
tion efficiency of 2009 pandemic H1N1. The paired mu-
tations S590 and R591, termed the SR polymorphism,
were present in only three of the 284 9 PB2 sequences of
human viruses before 2009 [23]. Other sites might affect
the polymerase activity as well. The mutations at posi-
tion 504 in PB2 (I504V) and position 550 in PA (I550L)
could result in enhanced virulence [24]. A special region
(residues 360-374) in the NP protein was found to play a
vital role in overcoming species barrier for 2009 pan-
demic H1N1 [25].
Compared to other proteins in the influenza viruses,
PB1-F2 is a newly discovered protein, which is unique
in that this protein is coded by a subset of the nucle-
otides that code for PB1 due to the use of a different
reading frame. The PB1-F1 protein has been implicated
in pathogenicity and the induction of cell death [26-28].
The 2009 pandemic H1N1 viru s has a truncated PB1-F1
protein, because its genome contains three stop codons
preventing PB1-F2 expression. Recently, studies found
that its function is not universal, but cell type and virus

586 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
strain dependent [29], and it plays a critical role in the
pathogenicity and tranmissibility of 2009 pandemic H1N1
[30,31].
NS1 is a multifunctional protein that contributes to
viral pathogenesis by neutralizing the interferon (IFN)
-based defense system of the host cell [32], and serves as
a strong inducer of apoptosis in infected human respira-
tory epithelial cells [33]. The NS1 protein of 2009 pan-
demic H1N1 is truncated and therefore missing a domain
responsible for increased pathogenicity of avian virus
[34]. There are two molecular markers that account for
the virulence of the highly pathogenic avian H5N1 vi-
ruses are not present in 2009 pandemic H1N1. They are
a lysine (K) at position 627 of PB2, and glutamic (E)
acid in position 92 of NS1 that might increase the repli-
cation efficiency and block host inhibition of viral repli-
cation, respectively [35,36].
A recent study on M gene identified sites of high se-
lective pressure between human and avian influenza,
which were 115, 121, 137 in M1, and 11, 16, 20, 54, 57,
78, 86, and 93 in M2 [37]. The 2009 pandemic H1N1
virus contains the adamantine-resistant mutation S31N
in its M2 protein, thus making the NA Inhibitors osel-
tamivir and zanamivir the only options available to treat
the infections caused by the pandemic virus [38].
Several studies focused on determining which amino
acid changes best disti ngui sh an avi an or swine influenza
virus from a human virus. An entropy analysis revealed
the human-avian host shift genomic signature of 52
markers in ten proteins of the influenza virus [39]. This
signature extended in [40] provided th e basis for finding
the amino acids of 2009 pandemic H1N1 at the host spe-
cies-specific positions to illustrate the adaptive muta-
tions of this virus. By comparison of the protein se-
quences of 2009 pandemic H1N1 with those in the pre-
vious pandemics and human, swine, and avian influenza
viruses, the mutation trend of the residues at the signa-
ture positions was discovered, and the potential roles of
the mutated residues in human adaptation and virulence
was probed in [41]. With mutual information analysis,
the characteristic sites for human-to -human transmission
in PB2 of influenza viruses were uncovered [42], and
subsequently a catalogue of 68 such sites in eight inter-
nal proteins were found to derive adaptation signatures
of viral proteomes [43], which included many of the 32
and 34 markers identified in [44,45], respectively.
Many of the molecular determinants associated with
adaptation to human hosts or to the emergence of a pan-
demic virus are not present in 2009 pandemic H1N1,
suggesting that other previously u nrecognized molecular
markers are responsible for its ability to infect humans
[17]. Therefore, uncovering new molecular features of
2009 pandemic H1N1 is of prime significance. In this
study, we collected all the protein sequences of the 2009
pandemic H1N1, 2009 seasonal H1N1, avian, human,
and swine influenza viruses available from the National
Center for Biotechnology Information (NCBI). Our ob-
jective was to explore the novel host markers in 2009
pandemic H1N1 that were not present in the classical
avian-human or sw ine-human ho st shift mark ers, and the
top markers that could differentiate 2009 pandemic
H1N1 from 2009 seasonal H1N1.
2. MATERIALS AND METHODS
2.1. Sequence Data
All influenza virus protein sequences were retrieved
from the Influenza Virus Resource (http://www.ncbi/
nlm.nih.giv/genomes/FLU/FLU.html) of the National
Center for Biotechnology Information (NCBI). Detailed
information about these sequences is in Table 1. All the
sequences used in the study were aligned with MAFFT
[46].
2.2. Random Forests
Random Forest, proposed by Leo Breiman in 1999 [47],
is an ensemble classifier based on many decision trees.
Each tree is built on a bootstrap sample from the origin al
training set and is unprun ed to obtain low-bias trees. The
variables used for splitting the tree nodes are a random
subset of the whole variable set. The classification deci-
sion of a new instance is made by majority voting over
all trees. About one-third of the instances are left of the
bootstrap sample and not used in the construction of the
tree. These instances in the training set are called “out-
of-bag” instances and are used to evaluate the perform-
ance of the classifier, which can achieve both low bias
and low variance with bagging and randomization.
2.3. Feature Selecti on Using Random For es ts
Random Forest calculates several measures of variable
importance. The mean decrease in accuracy measure was
employed in [48] to rank the importance of the features
in prediction. This measure is based on the decrease of
classification accuracy when values of a variable in a
node of a tree are permuted randomly. In this study, two
packages of R, randomForest and varSelRF [48], were
utilized to compute th e importance of the amino acids in
a given protein sequence dataset. The effectiveness and
robustness of this technique as a feature selection
method has been demonstrated in various studies [49-
54].
2.4. Procedure to Find Novel Host Sites in 2009
Pandemic H1N1
Four steps were created to locate the novel sites associated

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 587
Table 1. Counts of the influenza protein sequences used in the current study.
Host Subtype Protein Nu m b e r o f S eq u e n c e sYears Host SubtypeProteinNu m b e r o f S eq u e n c e sYears
Human Pandemic H1N1 HA 710 2009 HumanAll t y p e sN S1 509 Al l y ea r s
Human Pandemic H1N1 NA 643 2009 HumanAll t y p e sN S2 383 Al l y ea r s
Human Pandemic H1N1 NP 394 2009 HumanAl l t yp e sPA 279 Al l y e ar s
Human Pandemic H1N1 M1 490 2009 HumanAl l t yp e sPB1 289 All y e a r s
Human Pandemic H1N1 M2 482 2009 HumanAl l t yp e sPB2 269 All y e a r s
Human Pandemic H1N1 NS1 366 2009 Avian H1 HA 120 All years
Human Pandemic H1N1 NS2 358 2009 AvianN1 NA 1821 All years
Human Pandemic H1N1 PA 295 2009 AvianAl l t yp e sNP 2888 Al l y ea r s
Human Pandemic H1N1 PB1 31 1 2009 AvianA l l t y p e sM1 4232 All y e a rs
Human Pandemic H1N1 PB2 31 1 2009 AvianA l l t y p e sM2 3182 All y e a rs
Human Seasonal H1N1 HA 128 2009 AvianAll t y p e sN S1 4610 A ll y ea r s
Human Seasonal H1N1 NA 125 2009 AvianAll t y p e sN S2 3422 A ll y ea r s
Human Seasonal H1N1 NP 25 2009 Av ianAll t yp e sPA 3 106 Al l ye a r s
Human Seasonal H1N1 M1 129 2009 AvianAl l t y p e sPB1 2979 Al l y e a r s
Human Seasonal H1N1 M2 129 2009 AvianAl l t y p e sPB2 2643 Al l y e a r s
Human Seasonal H1N1 NS1 25 2009 SwineH1 HA 379 All years
Human Seasonal H1N1 NS2 25 2009 SwineN1 NA 278 All years
Human Seasonal H1N1 PA 23 2009 SwineAl l t y p esNP 4 20 A l l y e a rs
Human Seasonal H1N1 P B1 25 2009 S wineAl l ty p esM1 516 Al l y e a rs
Human Seasonal H1N1 P B2 25 2009 S wineAl l ty p esM2 406 Al l y e a rs
Human H1 HA 640 All yearsSwineAl l t y p e sN S1 506 Al l y e a rs
Human N1 NA 1 127 All yearsSwineAl l t y p esN S2 351 Al l y e a r s
Human All types NP 393 All yearsSwineAll typesPA 343 All years
Human All types M1 1512 All yearsSwineAll typesPB1 368 All years
Human All types M2 1415 All yearsSwineAll typesPB2 327 All years
with host adaptation in 2009 pandemic H1N1.
Step 1: For each protein, the consensus sequence of
avian, 2009 pandemic H1N1, human, and swine viruses
were calculated separately, and the positions with dif-
ferent amino acids of the four consensus sequences were
identified, since the different amino acids at these posi-
tions have the poten tial to contribute to host switches.
Step 2: For each protein, Random Forests were used
to identify the top 20 positions that have highest impor-
tance in separating avian from human viruses, and swine
from human viruses, respectively.
Step 3: Finding the intersection of the top positions
with importance larger than 0.005 for separating 2009
pandemic H1N1 from human viruses and the positions with
different co nsensus ami no acids found in step one.
Step 4: The positions discovered in step three minus
the positions found in step two will be the novel posi-
tions important for separatig 2009 pandemic H1N1 n

588 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
Table 2. This table contains the Hamming distances of consensus protein sequences of avian, human,
2009 pandemic H1N1, and swine viruses.
Proteins HA NANPM1M2NS1NS2PA PB1 PB2
Dist(Avian,2009_pandemic) 90 35 28 8 6 35 11 17 23 17
Dist(Human,2009_pandemic) 10961 42 19 15 42 13 30 21 31
Dist(Swine,2009_pandemic) 43 52 10 11 6 17 7 15 20 17
Dist(Avian,Human) 90 50 31 11 13 23 5 20 4 19
Dist(Avian,Swine) 61 37 21 3 2 22 7 4 6 2
Dist(Human,Swine) 97 52 35 8 11 36 7 19 6 21
from human viruses.
The purpose of steps 2 and 3 was to calculate the ad-
aptation signatures for various virus groups, which were
then used in step 4. The amino acids at most of these
novel sites in 2009 pandemic H1N1 turned out to be
different from those in avian, human, and swine viruses
that were the same at these positions.
Random Forests produce non-deterministic outcomes.
To compensate this bias, the Random Forests algorithm
was run multiple times and then the average of the results
was taken. The importance of each residue in the protein
sequences was based on the averaged calculations by
using the function randomVarImpsRF in varSelRF re-
peated 5 times.
3. RESULTS
3.1. Com p ar i s o n o f Cons e n s u s Prote i n Sequences
of Influenza Viruses
In considering the relationship among the proteins of
influenza viruses, the Hamming distance, defined as the
number of positions at which the corresponding amino
acids of two sequences are different, of any two consen-
sus protein sequences of avian, human, 2009 pandemic
H1N1, and swine viruses was calculated. The distance
information in Table 2 provided insight into the se-
quence similarity between the proteins of 2009 pan-
demic H1N1 and those of other virus groups. In particu-
lar, the distances between 2009 pandemic H1N1 and
avian, human, and swine viruses reflected the origin of
2009 pandemic H1N1 [2].
3.2. Novel Host Sites in the Proteins of 2009
Pandemic H1N1
Our analysis discovered a catalogue of novel host mark-
ers in the proteins of 2009 pandemic H1N1 that included
18 markers in HA, 15 in NA, 9 in PB2, 11 in PB1, 13 in
PA, 10 in NS1, 1 in NS2, 11 in NP, 3 in M1, and 1 in
M2. In the following sections, each of the ten proteins of
2009 pandemic H1N1 was compared to that of avian,
human, and swine viruses. Random Forests were em-
ployed to identify the top impo rtant positions in the pro-
teins of influenza that could separate 2009 pandemic
H1N1 from avian, human, and swine viruses, and the top
positions that could discriminate 2009 pandemic H1N1
and 2009 seasonal H1N1.
The novel host markers in 200 9 pandemic H1N1 were
uncovered with the procedure outlined in Section 2.3.
Some of the markers that could classify 2009 pandemic
H1N1 and 2009 seasonal H1N1 were also part of the
previously discovered genomic signature that separate
avian or swine from human viruses. Because the se-
quences of 2009 seasonal H1N1 were a subset of those
of human viruses, there were common important sites in
each protein between the sites in 2009 pandemic versus
2009 seasonal and the sites in 2009 pandemic versus
human viruses.
To render a complete picture of host shift markers of
different types, the novel sites in each of the ten proteins
of 2009 pandemic H1N1 were exhibited along with the
avian-human and swine-human sites in a single table.
The conservation of residues comprising these sites in
each protein as represented by their frequency at these
positions was also displayed in the table. The top impor-
tant sites in each protein for differentiating 2009 pan-
demic H1N1 from avian, human, and swine viruses were
displayed in a single figure, which were used in the pro-
cedure to find the novel sites.
Due to high genetic variation of the HA and NA pro-
teins, only the HA protein sequences of H1 subtype and
the NA protein sequences of N1 subtype of avian, hu-
man, and swine viruses were used to compare those of
2009 pandemic H1N1 in the current analysis. Therefore,
the novel markers in HA and NA of 2009 pandemic
H1N1 found in this study were subtype-specific. Be-
cause all the PB1-F2 proteins of 2009 pandemic H1N1
were truncated and nonfunctional, they were excluded in
this study.
3.2.1. HA Prot ein
As the primary target of host immune responses, the
surface protein HA is under high selection pressure, as
evidenced by the large number of amino acid substitu-
tions in this protein. There was a clear distinction of
amino acids at position 127, where the human HA had a

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 589
Table 3. This table contains the consensus amino acids and their frequency at positions in HA that have high importance in separat-
ing 2009 pandemic H1N1 from human H1 viruses. The single letter ‘a’ (for avian) or ‘s’ (for swine) in parenthesis after a position
number indicates whether the same position is also important for separating 2009 pandemic H1N1 from avian or swine viruses or
both. The novel host sites in this protein are the positions without an ‘a’ or a ‘s’ or both.
Position 71 84 120(a) 127(a,s) 128(s) 129(s) 130(a,s) 142 168 216 239 250
Avian L(93.3%) N(99.2%) A(96.7%) E(97.5%) T(100%) T(93.3%) K(94.2%)S(92.5%)N(99.2%) A(94.2%) T(97.5%) A(94.2%)
Human I(92.5%) N(98.4%) E(95.0%) -(100%) T(95.8%)V(95.0%) T(97.8%)S(88.4%)N(98.4%) K(96.1%) T(99.2%) A(100%)
2009
H1N1 S(100%) S(100%) T(100%) D(99.7%) S(97.0%)N(100%) K(100%) K(100%) D(100%)I(99.9%) K(100%) V(99.9%)
Swine F(58.6%) N(91.8%) A(44.9%) E(57.3%) T(81.0%)N(64.12%)R(62.0%)N(66.2%)N(85.5%) A(46.7%) T(80.0%) V(62.5%)
Position 257 258 260 261 298 302 314 365 374 493 527
Avian L(85.8%) N(95.0%) G(94.2%) S(98.3%) I(93.3%) E(98.33%)M(99.2%)Q(94.2%)G(100%)S(96.7%) L(99.2%)
Human L(96.7%) S(96.7%) G(98.0%) F(97.3%) V(99.5%)E(100%) M(99.5%)Q(98.9%)G(99.5%) S(99.2%) L(99.1%)
2009
H1N1 M(100%) E(99.9%) N(99.0%) A(99.6%) I(100%) K(100%) L(99.9%)L(100%) E(99.4%)A(100%) V(99.9%)
Swine M(54.1%) N(41.7%) G(78.1%) S(72.0%) V(73.1%)E(94.2%) M(95.3%)Q(61.7%)G(90.8%) S(67.0%) L(78.1%)
Table 4. This table contains the consensus amino acids and their frequency at positions in NA that have high importance in separating
2009 pandemic H1N1 from human N1 viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’
(for swine) or both.
Position 84 126 149 163(s) 166 189 257 269 285(s) 321
Avian T(76.4%) H(98.9%) V(95.2%) V(94.1%) A(99.6%) S(93.5%) K(96.8%) L(99.7%) A(97.5%) V(94.6%)
Human T(92.8%) H(100%) V(95.6%) L(83.8%) A(99.8%) G(86.4%) K(99.7%) L(100%) T(85.5%) V(99.9%)
2009 H1N1 K(100%) P(100%) I(100%) I(100%) V(99.7%) N(99.8%) R(99.8%) M(99.7%) S(100%) I(100%)
Swine I(55.8%) H(89.2%) V(73.0%) I(84.5%) A(64.0%) G(58.3%) K(90.3%) L(90.7%) T(55.04%) V(71.6%)
Position 331 365(a,s) 369(a,s) 385 389 395 397 398 436
Avian G(99.7%) T(91.4%) S(99.4%) S(88.3%) V(90.4%) A(99.2%) T(99.2%) D(99.5%) T(99.6%)
Human G(98.8%) N(84.7%) K(84.5%) S(99.7%) V(94.5%) A(99.7%) T(99.7%) D(99.7%) T(99.5%)
2009 H1N1 K(100%) I(99.8%) N(100%) N(100%) I(100%) G(100%) N(99.7%) E(100%) -(100%)
Swine G(68.7%) I(63.0%) S(82.0%) S(69.4%) V(38.1%) A(63.3%) T(87.8%) D(92.1%) T(99.6%)
deletion whereas the other three virus groups had not
(Table 3). However, as will be seen in the NA protein
section below (Table 4), the NA protein of 2009 pan-
demic H1N1 had a deletion at position 436 though the
other three virus groups had not. The positions in Table
3 including 71, 84, 130, 257, 258, and 314 had signifi-
cant effects on the receptor binding specificity of HA of
2009 pandemic H1N1[13]. HA has two functional do-
mains HA1 (residues 1-327) and HA2 (residues 328-
549). Evidently, most of the sites in Table 3 were in
HA1, illustrating a much higher selection pressure of
HA1 relative to HA2. The HA active site located in a
cleft is composed of the residues 91, 150, 152, 180, 187,
191, and 192. The active site cleft of HA is formed by its
right edge (131_GVTAA) and left edge (221_RGQAGR)
[55]. Four sites 127, 128, 129, and 130 in Table 3 were
near the right edge of the active site (Table 3).
3.2.2. NA Protein
In addition to the surface protein HA, the influenza A
virus also has NA as another surface protein, and the
balanced interplay between them is essential for the life
cycle of this virus. Because of its critical role in viral
replication and its highly conserved active sites, NA is
the main target for drug design against influenza virus.
The NA Inhibitors oseltamivir and zanamivir were the
only drugs available to treat the infections caused by
2009 pandemic H1N1, because the novel virus had an
adamantine-resistant mutation S31N in its M2 protein
[38]. As a result, the surveillance of any potential drug-
resistant mutations in the NA protein of 2009 pandemic
H1N1 received high priority. The mutation H275Y (N1
numbering, H274Y in N2 numbering) in NA is well
known for its resistance to NA Inhibitors. There were
123 Ys and 2 Hs in 125 NA sequences of 2009 seasonal
H1N1 and 12 Ys and 631 Hs in 643 NA sequences of
2009 pandemic H1N1 used in the current study. Both
NAs in 2009 pan demic and 2009 season al H1N1 did not
have the novel NA mutation Q136K [41] that confers
zanamivir resistance.
NA is also is constantly evolving under host immune
pressure, and the mutations in Table 4 illustrated its ge-
netic variation. As mentioned in the HA protein section
above, the NA of 2009 pandemic H1N1 had a deletion at

590 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
Figure 1. Top important HA positions in distinguishing avian H1, human H1, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine
H1 viruses.
Figure 2. Top important NA positions in distinguishing avian N1, human N1, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine
N1 viruses.
position 436 while the other three virus groups had not.
However, the HA of human virus had a deletion at posi-
tion 127 but the other three virus groups had not.
The NA active site is a shallow pocket constructed
from conserved residues, some of which contact the sub-
strate directly and participate in catalysis, while others
provide a structural framework [56]. According to the
numbering in [57], these residues of N1 are 118, 119,
JBiSE

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 591
151, 152, 156, 179, 180, 223, 225, 228, 247, 277, 278,
293, 295, 368, and 402. The antigenic sites of N1 are
residues 83-143, 156-190, 252-303, 330, 332, 340-345,
368, 370,38 7- 3 95 , 43 1-435, 448-468. Novel host sit es 84,
126, 166, 189, 257, 269, 389, and 395 were at the anti-
genic sites of N1 (Table 4).
Because of a common deletion in the stalk region of
the NA proteins of avian viruses, only the residues after
position 82 were includ ed in the Random Forest analysis
on avian, human, 2009 pandemic H1N1, and swine vi-
ruses. However, the whole NA sequences were used in
the analysis of 2009 pandemic and 2009 seasonal H1N1.
3.2.3. M1 Protein
M1 protein forms a shell inside the viral envelope to
offer strength and rigidity to the viral structure. M1 in-
teracts with HA, NA, M2, and lipid membranes during
budding of new virions from the cell surface, and func-
tions in the formation of vRNP complexe and the disso-
ciation of vRNP from the nuclear matrix, and in assem-
bly by recruiting the viral components to the site of as-
sembly. The dissociation of M1 from vRNP is triggered
by transport of hydrogen ions across the viral membrane
by M2, an early step preceding entry of vRNPs into the
cytoplasm of the host cells. M1 also binds to NS2 to
facilitate nuclear export of the vRNP [37]. There was a
mutation R101K in the M1 protein of 2009 pandemic
H1N1 (Table 5). It would be of interest to exam the im-
pact of this mutation on viral replication. The basic
amino acids 101RKLKR105 of M1 were involved in
vRNP binding and nuclear localization. In [58], the func-
tions of 101RKLKR105 were studied by introducing
mutations into the M gene of influenza virus A/WSN/33.
Individual substitution, R101S or R105S, had a minimal
effect on viral replication, but the double mutation
R101S-R105S reduced viral replication at a restrictive
temperature.
The M1 is a highly conserved protein. Therefore, the
changes of M1 may reflect host-specific adaptation. Po-
sitions 115, 121, and 137 were identified as avian-human
host shift markers in [43]. Our investigation indicated
position 218 was as important as position 121. Position
137 was a swine-human marker in [40], but our study
also revealed positions 115 and 218 were as important as
positions 137 as swine-human markers (Figure 3). The
novel site 30 was in the membrane binding domain [43],
and sites 207 and 209 were in the C-terminal part of M1
(residues 165-252) that binds to vRNP [59].
3.2.4. M2 Protein
This 97 amino acid-long integral membrane protein has
three domains, one N-terminal extracellular domain (24
residues) recognized by host immune system, one 19-
residue transmembrane domain responsible for ion
channel activity, and one 54-residue cytoplasmic tail
interacting with M1 and required for genome packing
and formation of virus particles [37]. Two M2 inhibitors
(adamantine and rimantadine) affect two steps in the
replication cycle, viral uncoating and viral maturation.
There are five known adamantine-resistanant mutations
in M2 (L26F, V27A, A30V, A30T, S31N, and G34E).
The 2009 pandemic H1N1 virus contains a mutation
S31N. They also contain a mutation L43T in M2 (Table
6 and Figure 4), which is not present in seasonal, tri-
ple-reassortant swine or H5N1 influenza viruses [15].
The replacement of the non-polar residue L43 by the
polar residue T43 in M2 may influence a nearby func-
tional residue W44, the channel lock and the binding site
of rimantadine [60]. Positions 11, 14, 20, 28, 54, 55, 57,
78, and 86 were avian-human host shift sites found in
[43]. However, the positions 18, 50, 86, and 93 were as
important as these sites in our examination. Positions 57 ,
86, and 93 were swine-human shift markers in [40], but
our analysis also included positions 28, 54, 77, 78, 79,
and 89 as swine-human markers with high importance
(Figure 4). The only novel site in this protein was 13
which was in the extracellular domain (Table 6).
3.2.5. NP Protein
The NP protein of the influenza virus binds the RNA
genome and functions as an adaptor between the virus
and the host cell. The interaction of the NP protein with
the viral polymerase is required for viral RNA replica-
tion, but not for the synthesis of viral messenger
RNAs(transcription). Previous experiments implicated
Table 5. This table contains the consensus amino acids and their frequency at positions in M1 that have high importance in separat-
ing 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’
(for swine) or both.
Position 15(a) 30 101(a,s) 115(a,s) 116(s) 121(a,s) 137(a,s) 142(a,s) 166(a,s) 207 209 214(s) 218(a,s)
Avian I(52.4%) D(99.9%) R(52.7%) V(99.5%) A(97.5%) T(96.2%) T(99.4%) V(91.6%)V(53.7%)S(68.4%) A(98.9%) Q(99.2%)T(99.8%)
Human V(67.6%) D(99.7%) R(93.1%) I(92.3%) A(98.7%) A(92.9%) A(93.1%) V(68.2%)V(93.2%)S(92.7%) A(99.9%) Q(99.8%)A(84.5%)
2009
H1N1 I(100%) S (99.8%) K(100%) V(99.4%) S(100%) T(100%) T(100%) A(99.6%)A(100%) N(99.8%) T(100%) H(100%) T(100%)
Swine V(65.3%) D(77.3%) R(64.7%) V(90.9%) A(68.4%) A(59.7%) T(93.2%) V(77.3%)V(66.1%)S(91.9%) A(78.5%) Q(67.1%)T(92.4%)

592 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
0.03
0.025
0.02
0.015
0.01
0.005
0
Importance
Figur e 3 . Top important M1 positions in dis tinguis hing avian , human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.
Table 6. This table contains the consensus amino acids and their frequency at positions in M2 that have high importance in separating
2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’ (for
swine) or both.
Position 11(a,s) 13 16(a,s) 20(a,s) 28(a,s) 31(s) 43(a,s) 77(s) 78(a,s) 86(a,s)
Avian T(92.0%) N(91.70%) E(92.4%) S(96.2%) I(54.5%) S(88.6%) L(97.0%) R(98.1%) Q(99.4%) V(99.6%)
Human I(91.6%) N(99.29%) G(91.2%) N(92.2%) V(96.8%) S(67.6%) L(67.7%) R(99.6%) K(60.9%) A(91.7%)
2009 H1N1 T(100%) S(99.79%) E(100%) S(100%) I(100%) N(100% ) T(100%) Q(100%) Q(100%) V(100%)
Swine T(52.7%) N(72.66%) E(57.9%) N(52.7%) I(38.9%) S(60.8%) L(93.8%) R(63.5%) Q(95.6%) V(94.6%)

Figure 4. Top important M2 positions in distinguishing avian, human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.
JBiSE

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 593
three NP r egions (residues 1-160, 2 56-340 and 340-49 8)
in binding to PB1 and PB2 [61]. Novel sites 21, 53, 119,
316, 353, 371, 377, 433, 444, and 498 were scattered in
these three regions (Table 7). One region, residues 360-
374, in NP of 2009 pandemic H1N1 was deemed ex-
tremely important for host range restriction, and is a
common feature of pandemic viruses [25]. Two posi-
tions 371 and 373 in Table 7 were in this region. Residue
100 was involved in the NP-PB2 interaction [62], and
ranked second in separating 2009 pandemic H1N1 from
avian viruses. The consensus amino acids at 100 of PB2
proteins of avian, human, 2009 pandemic H1N1, and
swine viruses were R, V, I and V respectively. The mu-
tation V100I might contribute to the increased transmis-
sibility or infection of 2009 pandemic H1N1 [41].
Positions 16, 33, 61, 100, 136, 214, 283, 305, 313,
357, 375, and 423 were avian-human host shift markers
in [43]. Furthermore, we found positions 31, 217, 373,
and 455 significant for discriminating avian and human
viruses (Figure 5).
3.2.6. NS1 Protein
All of the proteins in influenza virus are structural ex-
cept for NS1 and PB1-F2. This protein is designated as
non-structural because it is synthesized in infected cells,
but is not incorporated into virions. NS1 is a multifunc-
tional protein involved in both protein-protein and pro-
tein-RNA interactions. Its N-terminal region has an
RNA-binding domain (residues 1-73) and its C-terminal
region (residues 74-237) contains the effector domain
that inhibits the maturation and exportation of the host
cellular antiviral mRNAs [63].
Because of a truncation in the NS1 proteins of 2009
pandemic H1N1, only the first 219 residues of the NS1
proteins were included in our analysis. Positions 22, 60,
81, 84, 215, and 227 were avian-human host shift sites in
[43], whereas our Random Forests analysis implied po-
sitions 79, 81, 114, 171, and 215 were as significant as
Table 7. This table contains the consensus amino acids and their frequency at positions in NP that have high importance in separating
2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’ (for
swine) or both.
Position 21 31(a,s) 53 119 189(s) 190 217(a) 289(s) 313(a,s) 316
Avian N(99.20%) R(99.8%) E(99.90%) I(97.65%) M(99.1%)V(98.61%)I(94.8%) Y(99.2%) F(99.0%) I(99.58%)
Human N(97.5%) K(65.4%) E(100%) I(97.46%) M(97.5%)V(97.20%)S(48.1%) Y(97.5%) Y(78.1%) I(99.75%)
2009 H1N1 D(100%) R(100% ) D(100%) V(100%) I(99.75%) A(100%) V(98.2%) H(99.7%) V(100%) M(100%)
Swine D(61.19%) R(82.6%) E(98.10%) V(57.14%)I(60.00%) A(56.42%)I(76.90%) H(64.29%) F(86.66%) I(95.24%)
Position 350(s) 353 371 373(a) 377 430(s) 433 444 456(s) 498
Avian T(94.39%) V(90.30%)M(94.77%) T(69.18%)S(69.67%)T(94.8%) T(95.36%)I(99.00%) V(98.4%) N(96.09%)
Human T(97.20%) S(52.42%)M(91.35%) A(34.35%)S(80.66%)T(83.72%)T(88.04%)I(98.22%) V(82.95%) N(96.18%)
2009 H1N1 K(100%) I(99.75%) V(100%) T(76.40%)N(100%) S(100%) N(100%) V(100%) L(100%) S(99.24%)
Swine K(64.52%) V(53.81%)V(59.29%) A(57.38%)S(51.90%)S(38.57%)N(60.95%)I(64.76%) L(61.19%) N(69.29%)
Figure 5. Top important NP positions in distinguishing avian, human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.
JBiSE

594 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
Table 8. This table contains the consensus amino acids and their frequency at positions in NS1 that have high importance in separat-
ing 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’
(for swine) or both.
Position 6 25(s) 59 67(s) 74 76 91 111 112(a)
Avian V(77.85%) Q(79.35%) R(73.41%) R(78.85%)D(97.87%)A(77.79%)T(97.33%) V(79.76%) A(59.11%)
Human V(95.48%) Q(95.09%) H(45.19%) R(54.62%)D(97.64%)A(97.84%)T(96.66%) V(96.07%) E(55.00%)
2009 H1N1 M(99.73%) N(99.73%) L(100%) W(100%) S(99.73%)T(1 00 % ) S(99.73%) I(100%) I(99.45%)
Swine V(95.26%) N(58.70%) L(59.49%) W(59.68%)D(37.49%)T(58.70%)A(60.28%) V(50.59%) A(36.96%)
Position 119 129(s) 171(a,s) 198 205 206(s) 207 213(s) 217(s)
Avian M(99.39%) I(75.55%) D(48.87%) L(52.52%)S(69.50%)S(67.66%)D(59.20%) P(92.56%) K(68.00%)
Human M(85.85%) M(51.28%) I(55.20%) L(84.28%)S(92.93%)S(91.55%)N(78.00%) P(97.05%) K(69.16%)
2009 H1N1 L(100%) V(100%) Y(100%) I(100%) N(100%) C(100%) D(100%) S(100%) E(100%)
Swine M(90.12%) I(64.23%) D(59.29%) L(97.04%)S(64.62%)R(56.13%)N(92.09%) P(51.38%) E(57.51%)

Figure 6. Top important NS1 positions in distinguishing avian, human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.

these sites (Figure 6). There were two novel sites 6 and
59 in the RNA-binding domain and the other novel sites
in the effector domain (Table 8).
3.2.7. NS2 Protein
Influenza virus replicates its RNA genome in the nucleus
of infected cells. The NS2 protein mediates the nuclear
export of virion RNAs, with help from M1 and NP. A
recent report indicated that it also has a role in the regu-
lation of viral transcription and replication [64]. NS2
contains a highly conserved nuclear export signal motif
in its amino-terminal region (residues 12-21) [65], and
site 14 in Table 9 was in this region.
Positions 60, 70, and 107 were avian-human ho st shift
markers in [43]. We found position 14 important as a
host marker as well. Position 107 was a swine-human
host switch marker in [40], but our analysis also pointed
to positions 14, 32, 49, and 57 as such sites of high sig-
nificance (Figure 7). The NS2 protein of 2009 pandemic
H1N1 contained so many important avian-human or swine-
human sites, resulting in only one site as a novel site
(Table 9).
3.2.8. PA Protein
Compared to the well-defined functions of PB1 and PB1,
PA is involved in a diverse range of functions of the po-
lymerase complex, including protein stability, endonu-
clease activity, and cap binding and promoter binding

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 595
[66]. Positions 28, 55, 57, 65, 66, 100, 225, 268, 321,
337, 356, 382, 400, 404, 409, 421, and 552 were
avian-human host shift markers in [43]. Additionally, we
found positions 241 and 383 equally important as these
positions as avian-hu man markers. Positions 268 and 552
were swine-human markers uncovered in [40]. Our
analysis suggested the positions 28, 225, 337, and 400
were equally crucial as these two sites as swine-human
markers (Figure 8).
The N-terminal domain of PA (residues 1-256) har-
bors several functional domains, including an endonu-
clease active site with a putative active site motif, two
putative nuclear transport motifs (residues 124-139
(NLS1) and residues 186-247 (NLS2)), and a proteolytic
domain that can induce generalized proteolysis of both
viral and host proteins. The C-terminal domain of PA
(residues 257-716) binds to PB1 for complex formation
and nuclear transport [66]. There were three novel sites
186, 204, and 213 within the second putative nuclear
localization signals (NLS2), and one novel site 626
within the PB1 binding domain (Table 10).
3.2.9. P B1 Protein
The influenza virus polymerase is responsible for repli-
cation and transcription of the eight gene segments of the
viral RNA genome in the infected host cell. PB1 can
interact with PB2, PA, and NP and binds to viral pro-
moter, and is accountable for viral RNA elongation and
cap RNA cleavage activities [66,67].
Position 336 was the only avian-human host shift
markers in [43]. We found positions 212, 327, 361, 375,
384, 401, 473, and 584 equally significant as position
336 (Figure 9). There were one novel site 12 within the
PB1-PA binding domain (residues 1-25) and two novel
sites 618 and 728 in the PB1-PB2 binding domain (resi-
dues 600- 7 57 ) (Table 11) [68].
PB1-PA binding domain (residues 1-25) and two novel
Table 9. This table contains the consensus amino acids and their frequency at positions in NS2 that have high importance in separat-
ing 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’
(for swine) or both.
Position 6 14(a,s) 32(s) 34(s) 40(a,s) 48(a) 57(a,s)
Avian V(79.1%) M(56.2%) I(99.0%) Q(95.6%) L(71.54%) A(73.2%) S(97.6%)
Human V(98.2%) L(55.6%) I(99.0%) Q(98.7%) L(61.9%) A(96.6%) S(60.3%)
2009 H1N1 M(99.7%) M(100%) V(99.7%) R(100%) I(100%) T(100%) Y(100%)
Swine V(95.2%) M(83.2%) V(68.1%) Q(53.3%) I(67.8%) A(70.1%) Y(65.8%)
Position 60(a,s) 63(a,s) 83(a) 89(a,s) 107(a,s) 115(a)
Avian S(55.7%) G(75.8%) V(71.7%) I(69.8%) L(99.9%) T(84.4%)
Human N(68.7%) G(96.3%) V(98.2%) T(63.2%) F(74.9%) T(89.8%)
2009 H1N1 S(100%) E(95.8%) M(99.7%) A(97.8%) L(100%) A(99.4%)
Swine N(61.0%) E(62.1%) V(97.4%) M(32.5%) L(90.9%) T(98.3%)

Figure 7. Top important NS2 positions in distinguishing avian, human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.

596 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
Table 10. This table contains the consensus amino acids and their frequency at positions in PA that have high importance in separat-
ing 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a ‘s’
(for swine) or both.
Position 28(a,s) 55(a) 85 100(a,s) 186 204 213 256 262 275 277
Avian P(99.5%) D(98.9%)T(96.5%) V(94.5%)G(98.0%)R(80.6%)R(97.8%)R(98.6%)K(96.9%) P(97.5%)S(97.1%)
Human L(70.3%) N(71.3%)T(91.0%) A(70.6%)G(100%) R(53.0%)R(97.5%)R(54.5%)K(95.0%) P(98.9%)S(35.8%)
2009 H1N1 P(100%) D(100%) I(100%) V( 100%) S(100%) K(100%)K(100%) K(100%) R(100%) L(98.6%)H(100%)
Swine P(79.9%) D(53.4%)T(60.9%) V(86.6%)G(96.8%)R(90.7%)R(92.1%)R(63.8%)K(77.8%) P(91.5%)S(53.4%)
Position 336 337(a,s) 356(a) 362 388 400(a,s) 404(a,s) 407 552(a,s) 626
Avian L(99.5%) A(88.7%)K(98.9%) K(99.5%)S(80.1%)S(40.9%)A(93.2%)I(95.9%) T(99.7%) K(83.7%)
Human L(97.1%) S(35.8%) R(69.9%) K(98.6%)S(84.2%)L(79.2%)S(72.0%)I(98.6%) S(71.7%) K(98.2%)
2009 H1N1 M(100%) A(99.0%)R(99.7%) R(100%) G(99.0%)P(100%) A(100%) V(99.3%)T(99.7%) R(99.3%)
Swine L(95.6%) A(86.9%)K(53.9%) K(71.1%)S(52.8%)P(30.6%)A(86.9%)I(71.1%) T(91.5%) K(96.2%)

Figure 8 . Top important PA positio ns in distinguishing avian, human , 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.
sites 618 and 728 in the PB1-PB2 binding domain (resi-
dues 600- 7 57 ) (Table 11) [68].
3.2.10. PB2 Protein
PB2 interacts with PB1 and NP, but not PA. Its primary
function is binding to cap structures on host cell pre-
mRNAs before they are cleaved to provide primers for
viral mRNA synthesis [66]. Positions 9, 44, 64, 81, 105,
199, 271, 292, 368, 475, 567, 588, 613, 627, 661, 674,
and 702 were avian-human host shift markers in [43].
Positions 108, 197, and 684 were as significant as these
sites in our finding. Position 44 was a swine- human
marker in [40], but our analysis i mplied positions 64, 65,
81, 105, 199, 292, 567, 627, 649, 661, and 674 were
equally important as position 44 (Figure 10). Position
702, an avian-human marker selected in [40,43,69],
ranked 21th in our Random Forests analysis, and there-
fore it was not included in our plot in Figure 10. In addi-
tion to the SR polymorphism, S590 and R591, found in
[23], novel sites in PB2 discovered here provided addi-
tional polymorphism that might convey enhanced poly-
merase activity in human cells. Position 627 in PB2 was
considered critical for host shifts in our analysis, a
well-known host marker discussed in [21,22], and was
located in the PB2-PB1 and PB2-NP b inding domains [43].
The PB2-NP binding domain contains residues 1-269
and 580-683, and the PB2-PB1 binding domain contains
residues 51-259 and 580-759. There w ere novel sites 54,
590, 645, and 667 in the PB2-PB1 and PB2-NP binding

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 597
Table 1 1 . This table contains the consensus amino acids and their frequency at positions in PB1 that have high importance in sepa-
rating 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a
‘s’ (for swine) or both.
Position 12 175 179 216 298 327(a,s) 339(s) 361(a,s) 364 386(s)
Avian V(99.0%) D(95.9%) M(95.9%) S(95.7%) L(98.6%) R(98.7%) I(98.7%) S(99.0%) L(99.5%) R(57.0%)
Human V(99.7%) D(97.2%) M(69.6%) S(65.7%) L(79.2%) K(53.6%) I(96.9%) S(61.9%) L(99.3%) R(65.4%)
2009 H1N1 I(99.7%) N(100%) I(100%) G(100%) I(100%) R(100%) M(100%) R(100%) I(100%) K(100%)
Swine V(95.1%) D(96.7%) M(66.0%) S(60.1%) L(96.7%) R(89.7%) I(44.6%) N(32.3%) L(96.5%) R(95.4%)
Position 435 486 517(s) 584(a,s) 587 618 638(s) 728 741(a,s)
Avian T(99.0%) R(98.6%) I(99.1%) R(97.1%) A(98.4%) E(97.7%) E(98.8%) I(99.1%) A(96.2%)
Human T(99.3%) R(64.7%) I(81.3%) R(63.7%) A(98.6%) E(99.7%) E(98.6%) I(100%) A(59.2%)
2009 H1N1 I(99.4%) K(100%) V(100%) Q(100%) V(97.4%) D(100%) D(100%) V(10 0%) S(100%)
Swine T(66.8%) R(64.1%) I(75.5%) R(40.5%) A(86.7%) E(61.4%) E(69.0%) I(98.4%) A(59.0%)


Figur e 9 . Top important PB1 pos itions in d istinguis hing avian , human, 2 009 pandemic H1N1, 2009 s easonal H1N 1, and swin e viruses .
domains and sites 147 and 225 in the PB2-NP binding
domain [43] (Table 12).
4. DISCUSSIONS
Extensive research to date provided highly informative
knowledge about th e origin and genetic lin eages of 2009
pandemic H1N1, but the host markers of this new virus
remained elusive. Recent studies indicated that human
host adaptation is complex and multigenic, and the
well-known host shift markers are lacking in this new
virus. The hypothesis in the current study was that these
markers of 2009 pandemic H1N1 might exist outside of
the space of traditional host switch markers. To test this
hypothesis in this study, Rando m Forests were applied to
uncover novel important markers in each of the ten pro-
teins of influenza that could differentiate 2009 p andemic
H1N1 from human viruses, but were not present in the
previous avian-human or swine-human host switch mark -
ers.
Our approach naturally led to a systematic discovery
of new host markers like the SQ polymorphis m found in
[23] that could enrich our current knowledge of 2009
pandemic H1N1 and complement the repertoire of ex-
isting host shift signatures. Among others, this study
revealed th e n ovel ho st sites 54, 14 7, 22 5, 315 , 45 3, 55 9,
590, 645, and 667 in PB2 of 2009 pandemic H1N1.
They provided ample potential sites to investigate ex-
perimentally whether they also compensate the lack of

598 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
amino acid lysine at residue 627, as the SR polymor-
phism. Their prospective broader roles in enhancing this
new virus’s replication and transmission in humans are
worthy of further research. In this regard, the three posi-
tions 54, 315, and 559 in PB2 were particularly of inter-
est because they had much higher importance than the
two positions 590 and 591 associated with the SR poly-
morphism.
Four proteins are involved and required in the synthe-
sis of influenza virus RNA, which are PB2 and PA of
avian lineage, PB1 of human origin, and NP derived
from classical swine viruses in 2009 pandemic H1N1.
To gain insight into the adaptive strategies employed by
these four proteins of different origins to evade restric-
tion in human cells will be a challenge. The novel sites
identified in this study provided a starting point for fu-
ture integrative examination of the interactions of these
proteins.
It was expected that 2009 pandemic H1N1 would co-
circulate with seasonal H1N1 for some time. Our cata-
logue of amino acid markers that could effectively sepa-
rate 2009 pandemic H1N1 from 2009 seasonal H1N1
presented a valuable view of these two viruses that share
similar clinical courses but are unique gen etically.
Table 12. This table contains the consensus amino acids and their frequency at positions in PB2 that have high importance in sepa-
rating 2009 pandemic H1N1 from human viruses. The novel host sites in this protein are the positions without an ‘a’ (for avian) or a
‘s’ (for swine) or both.
Position 9(a) 54 64(a,s) 65(s) 81(a,s) 105(a,s) 147 184(s) 199(a,s)
Avian D(97.5%) K(99.7%) M(74.9%) E(97.8%) T(97.3%) T(90.9%) I(82.1%) T(96.3%) A(99.2%)
Human N(71.0%) K(100%) T(68.0%) E(98.5%) M(52.0%) V(52.4%) I(87.7%) T(99.6%) S(72.9%)
2009 H1N1 D(10 0%) R(100%) M(100%) D(99.7%) T(100%) T(100%) T(100%) A(99.0%) A(100%)
Swine D(63.9%) K(98.5%) M(53.2%) E(69.4%) T(84.7%) T(87.8%) I(68.2%) T(57.2%) A(55.0%)
Position 225 292(a,s) 315 340(s) 453 475(a) 559 567(a,s) 588(a,s)
Avian S(99.4%) I(88.6%) M(95.2%) R(52.2%) P(94.7%) L(99.2%) T(91.0%) D(98.0%) A(95.8%)
Human S(98.9%) T(73.6%) M(99.6%) R(60.2%) H(52.0%) M(70.3%) T(71.7%) N(70.3%) I(68.4%)
2009 H1N1 G(100%) V(99.4%) I(100%) K(97.4%) S(99.7%) L(100%) I(100%) D(100%) T(98.4%)
Swine S(72.2%) I(56.0%) M(97.2%) R(59.9%) P(57.8%) L(54.1%) T(70.9%) D(90.2%) A(55.7%)
Position 590 591(s) 613(a,s) 627(a,s) 645 661(a,s) 667 674(a,s) 684(a)
Avian G(87.6%) Q(97.9%) V(98.4%) E(91.7%) M(99.4%) A(96.2%) V(92.4%) A(97.1%) A(96.9%)
Human G(69.9%) Q(98.9%) T(64.3%) K(80.3%) M(98.9%) T(78.4%) I(62.1%) T(69.1%) A(51.3%)
2009 H1N1 S(99.7%) R(100%) V(100%) E(100%) L(100%) A(100%) V(100%) A(99.7%) S(100%)
Swine G(71.3%) Q(67.3%) V(74.9%) E(53.8%) M(74.9%) A(48.0%) V(69.4%) A(86.5%) A(74.3%)

Figure 10. Top important PB2 positions in distinguishing avian, human, 2009 pandemic H1N1, 2009 seasonal H1N1, and swine viruses.

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 599
Various computational techniques including entropy
[39,40], mutual information [42,43], statistical tests [44 ],
and support vector machines [45] were utilized to dis-
cover molecular markers in influenza viruses. To dem-
onstrate the validity of using Random Forests as a fea-
ture selection technique in identifying novel host
markers in 2009 pandemic H1 N1, the top markers found
by Random Forests to distinguish the human virus from
avian or swine viruses were also included in this report,
which contained many known host adoption markers
from previous studies. There were fewer novel sites in
M1, M2, and NS2 than in the other proteins under this
study resulting from many avian-human or swine-human
sites among these proteins.
5. CONCLUSIONS
Our findings confirmed that there are novel host sites in
the proteins of 2009 pandemic H1N1 that could separate
this new virus from human viruses with high confidence.
These markers could not be found in the search space of
traditional avian-human or swine-human host shift mark-
ers, thus offering new potential sites for further experi-
mental verificat ion to el ucidate t heir bi ologi cal functi ons.
6. ACKNOWLEDGEMENTS
We thank Houghton College for its financial support.
REFERENCES
[1] Samji, T. (2009) Influenza A: Understanding the viral
life cycle . Yale Journal of Biology Medicine, 82(4), 153-
159.
[2] Gavin, J.D., Smith, D.V., Bahl, J., Lycett, S.J., et al.
(2009) Origins and evolutionary genomics of the 2009
swine-origin H1N1 influenza A epidemic. Nature, 459,
1122-1125.
[3] Chang, Y.S., van Hal, S.J., Spencer, P.M, Gosbell, I.B.
and Collett, P.W. (2010) Comparison of adult patients
hospitalised with pandemic (H1N1) 2009 influenza and
seasonal influenza during the PROTECT phase of the
pandemic response. The Medical Journal of Australia,
192(2), 90-93.
[4] Dhiman, N., Mark, J.E., Irish, C., Wright, P., Smith, T.F.
and Pritt, B.S. (2010) Mutability in the matrix gene of
novel influenza A H1N1 virus detected using a fret
probe-based real-time reverse transcriptase PCR assay.
Journal of Clinical Microbiology, 48(2), 677-679.
[5] Zheng, X., Todd, K.M., Yen-Lieberman, B., Kaul, K.,
Mangold, K. and Shulman, S.T. (2009) Unique finding of
a 2009 H1N1 influenza virus－positive clinical sample
suggests matrix gene sequence variation. Journal of
Clinical Microbiology, 48(2), 665-666.
[6] Shen, J., Ma, J. and Wang, Q. (2009) Evolutionary trends
of A (H1N1) influenza virus hemagglutinin since 1918.
PLoS One, 4(11), e7789.
[7] Soundararajan, V., Tharakaraman, K., Raman, R., Ragu-
ram, S., Shriver, Z., Sasisekharan, V. and Sasisekharan,
R. (2009) Extrapolating from sequence—the 2009 H1N1
‘swine’ influenza virus. Nature Biotechnology, 27, 510-
513.
[8] Childs, R.A., Palma, A.S. , Wharton, S., Matrosovich, T.,
Liu, Y., Chai, W.G., Campanero-Rhodes, M.A., et al.
(2009) Receptor-binding specificity of pandemic influ-
enza A (H1N1) 2009 virus determined by carbohydrate
microarray. Nature Biotechnology, 27, 797-799.
[9] Igarashi, M., Ito, K., Yoshida, R., Tomabechi, D., Kida,
H. and Takada, A. (2009) Predicting the antigenic struc-
ture of the pandemic (H1N1) 2009 influenza virus he-
magglutinin. PLoS One, 5(1), e8553.
[10] Cosic, I. (1997) The resonant recognition model of mac-
romolecular bioreactivity, theory and application. Birk-
hauser Verlag, Berlin.
[11] Veljkovic, V., Niman, H.L., Glisic, S., Veljkovic, N.,
Perovic, V. and Muller, C.P. (2009) Identification of he-
magglutinin structural domain and polymorphisms which
may modulate swine H1N1 interactions with human re-
ceptor. BMC Structural Biology, 9, 62.
[12] Hu, W. (2010) Identification of highly conserved do-
mains in hemagglutinin associated with the receptor
binding specificity of influenza viruses: 2009 H1N1, avian
H5N1 and swine. Journal of Biomedical Science and
Engineering, 3, 114-123.
[13] Hu, W. (2010) Quantifying the effects of mutations on
receptor binding specificity of influenza viruses. Journal
of Biomedical Science and Engineering, 3, 227-240.
[14] Janies, D.A., Voronkin, I.O., Studer, J., Hardman, J.,
Alexandrov, B.B., Treseder, T.W. and Valson, C. (2010)
Selection for resistance to oseltamivir in seasonal and
pandemic H1N1 influenza and widespread co-circulation
of the lineages. International Journal of Health Geogra-
phics, 9(1), 13.
[15] Deyde, V.M., Sheu, T.G., Trujillo, A.A., Okomo-
Adhiambo, M., Garten, R., Klimov, A.I. and Gubareva,
L.V. (2010) Detection of molecular markers of drug re-
sistance in 2009 pandemic influenza A (H1N1) viruses
by pyrosequencing. Antimicrob Agents Chemother, 54(3),
1102-1110.
[16] Hurt, A.C., Holien, J.K., Parker, M., Kelso, A. and Barr,
I.G. (2009) Zanamivir-resistant influenza viruses with a
novel neuraminidase mutation. The Journal of Virology,
83(20), 10366-10373.
[17] Garten, R.J., Davis, C.T., Russell, C.A., Shu, B., Lind-
strom, S., Balish, A., Sessions, W.M., Xu, X., et al. (2009)
Antigenic and genetic characteristics of swine-origin
2009 A(H1N1) influenza viruses circulating in humans.
Science, 325(5937), 197-201.
[18] Itoh, Y., Shinya, K., Kiso, M. , Watanabe, T., Sakoda, Y.,
Hatta, M., Muramoto, Y., et al. (2009) In vitro and in
vivo characterization of new swine-origin H1N1 influ-
enza viruses. Nature, 460, 1021-1025.
[19] Uhlendorff, J., Matrosovich, T., Klenk, H.D. and Ma-
trosovich, M. (2009) Functional significance of the he-
madsorption activity of influenza virus neuraminidase
and its alteration in pandemic viruses. Archives of Virol-
ogy, 154(6), 945-957.
[20] Sung, J.C., van Wynsberghe A.W., Amaro, R.E., Li,
W.W. and McCammon, J.A. (2010) Role of secondary

600 W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601
sialic acid binding sites in influenza N1 neuraminidase.
Journal of the American Chemistry Society, 132(9),
2883-2885.
[21] Steel, J., Lowen, A., Mubareka, S., Palese, P. and Baric,
R. (2009) Transmission of influenza virus in a mammal-
ian host is increased by PB2 amino acids 627K or
627E/701N. PLoS Pathog, 5, e1000252.
[22] Subbarao, E.K., London, W. and Murphy, B.R. (1993) A
single amino-acid in the Pb2-gene of influenza-A virus is
a determinant of host range. Journal of Virology, 67,
1761-1764.
[23] Mehle, A. and Doudna, J.A. (2009) Adaptive strategies
of the influenza virus polymerase for replication in hu-
mans. Proceedings of the National Academy of Sciences
of the United States of America, 106(50), 21312-21316.
[24] Rolling, T., Koerner, I., Zimmermann, P., Holz, K., Hal-
ler, O., Staeheli, P. and Kochs, G. (2009) Adaptive muta-
tions resulting in enhanced polymerase activity contrib-
ute to high virulence of influenza A virus in mice. Jour-
nal of Virology, 83 (13), 6673-6680.
[25] Liu, X. and Zhao, Y.P. (2010) Switch region for patho-
genic structural change in conformational disease and its
prediction. PLoS One, 5(1), e8441.
[26] Chen, W., Calvo, P.A., Malide, D., Gibbs, J., Schubert,
U., Bacik, I., Basta, S., O'Neill, R., Schickli, J., Palese, P.,
Henklein, P., Bennink, J.R. and Yewdell, J.W. (2001) A
novel influenza A virus mitochondrial protein that in-
duces cell death. Nature Medicine, 7, 1306-1312.
[27] Lamb, R.A. and Takeda, M. (2001) Death by influenza
virus protein. Nature Medici ne, 7, 1286-1288.
[28] Zell, R., Krumbholz, A., Eitner, A., Krieg, R., Halbhuber,
K.J. and Wutzler, P. (2007) Prevalence of PB1-F2 of in-
fluenza A viruses. Journal of General Virology, 88, 536-
546.
[29] McAuley, J.L., Zhang, K. and McCullers, J.A. (2010)
The effects of influenza A virus PB1-F2 protein on po-
lymerase activity are strain specific and do not impact
pathogenesis. Journal of Virology, 84(1), 558-564.
[30] Ramakrishnan, M.A., Gramer, M.R., Goyal, S.M. and
Sreevatsan, S. (2009) A Serine12Stop mutation in PB1-
F2 of the 2009 pandemic (H1N1) influenza A: A possible
reason for its enhanced transmission and pathogenicity to
humans. Journal of Veterinary Science, 10(4), 349-351.
[31] Trifonov, V. and Rabadan, R. (2009) The contribution of
the pb1-f2 protein to the fitness of influenza a viruses
and its recent evolution in the 2009 influenza A (H1N1)
pandemic virus. PLoS Current: Influenza, 21, RRN1006.
[32] Hale, B.G., Randall, R.E., Ortín, J. and Jackson, D. (2008)
The multifunctional NS1 protein of influenza A viruses,
Journal of General Vi rology, 89, 2359-2376.
[33] Zhang, C.F., Yang, Y.T., Zhou, X.W., Liu, X.L., Song,
H.B., He, Y.X. and Huang, P.T. (2010). Highly patho-
genic avian influenza A virus H5N1 NS1 protein induces
caspase-dependent apoptosis in human alveolar basal
epithelial cells. Virology Journal, 7, 51.
[34] Jackson, D., Hossain, M.J., Hickman, D., Perez, D.R. and
Lamb, R.A. (2008) A new infl uenza virus virulence de-
terminant: The NS1 protein four C-terminal residues
modulate pathogenicity. Proceedings of the National
Academy of Sciences of the United States of America,
105, 4381-4386.
[35] Seo, S.H., Hoffmann, E. and Webster, R.G. (2002) Le-
thal H5N1 influenza viruses escape host anti-viral cyto-
kine responses. Nature Medicine, 8, 950-954.
[36] Salomon, R., Franks, J., Govorkova, E.A., Ilyushina,
N.A., Yen, H.L., Hulse-Post, D.J., Humberd, J., Trichet,
M., Rehg, J.E., Webby, R.J., Webster, R.G. and Hoff-
mann, E. (2006) The polymerase complex genes contrib-
ute to the high virulence of the human H5N1 influenza
virus isolate A/Vietnam/1203/04, Journal of Experimen-
tal Medicine, 203(3), 689-697.
[37] Furuse, Y., Suzuki, A., Kamigaki, T. and Oshitani, H.
(2009) Evolution of the M gene of the influenza A virus
in different host species: Large-scale sequence analysis.
Virology Journal, 6, 67.
[38] Furuse, Y., Suzuki A. and Oshitani, H. (2009) Large-
scale sequence analysis of M gene of influenz a A viruse s
from different species: Mechanisms for emergence and
spread of amantadine resistance. Antimicrobial Agents
and Chemotherapy, 53(10), 4457-4463.
[39] Chen, G.W., Chang, S.C., Mok, C.K., Lo, Y.L., Kung,
Y.N., et al. (2006) Genomic signatures of human versus
avian influenza A viruses. Emerging Infectious Diseases,
12, 1353-1360.
[40] Chen, G.W. and Shih, S.R. (2009) Genomic signatures of
influenza A pandemic (H1N1) 2009, Virus. Emerging
Infectious Diseases, 15, 1897-1903.
[41] Pan, C., Cheung, B., Tan, S., Li, C., Li, L., et al. (2010)
Genomic signature and mutation trend analysis of pan-
demic (H1N1) 2009, Influenza A virus. PLoS One, 5(3),
e9549.
[42] Miotto, O., Heiny, A., Tan, T.W., August, J.T., Brusic, V.
(2008) Identification of human-to-human transmissibility
factors in PB2 proteins of influenza A by large-scale
mutual information analysis. BMC Bioinformatics, 9,
S18.
[43] Miotto, O., Heiny, A.T., Albrecht, R., García-Sastre, A.,
Tan, T.W., August, J.T. and Brusic, V. (2010) Com-
plete-proteome mapping of human influenza A adaptive
mutations: implications for human transmissibility of
zoonotic strains. PLoS One, 5(2), e9025.
[44] Finkelstein, D.B., Mukatira, S., Mehta, P.K., Obenauer,
J.C., Su, X., Webster, R.G. and Naeve, C.W. (2007) Per-
sistent host markers in pandemic and H5N1 influenza
viruses. Journal of Virology, 81(19), 10292-10299.
[45] Allen, J.E., Gardner, S.N., Vitalis, E.A., Slezak, T.R.
(2009) Conserved amino acid markers from past influ-
enza pandemic strains. BMC Microbioloy, 9, 77.
[46] Katoh, K., Kuma, K., Toh, H. and Miyata, T. (2005)
MAFFT version 5: Improvement in accuracy of multiple
sequence alignment. Nucleic Acids Research, 33, 511-
518.
[47] Breiman, L. (2001) Random Forests. Machine Learning,
45(1), 5-32.
[48] Díaz-Uriarte, R. and Alvarez de Andrés, S. (2006) Gene
selection and classification of microarray data using
random forest. BMC Bioinformatics, 7, 3.
[49] Archer, K.J. and Kimes, R.V. (2008) Empirical charac-
terization of random forest variable importance measures.
Computational Statistics and Data Analysis, 52, 2249-
2260.
[50] Reif, D.M. Motsinger, A.A., McKinney, B.A., Crowe,
J.E. and Moore, J.H. (2006) Feature selection using a
random forests classifier for the integrated analysis of

W. Hu / J. Biomedical Science and Engineering 3 (2010) 584-601 601
multiple data types. Proceedings of 2006 IEEE Sympo-
sium on Computational Intelligence and Bioinformatics
and Computational Biology, Toronto.
[51] Granittoa, P.M., Furlanellob, C., Biasiolia, F. and
Gasperia, F. (2006) Recursive feature elimination with
random forest for PTR-MS analysis of agroindustrial
products. Chemometrics and Intelligent Laboratory Sys-
tems, 83, 83-90.
[52] Menze1, B.H., Kelm, B.M., Masuch, R., Himmelreich,
U., Bachert, P., Petrich, W. and Hamprecht, F.A. (2009)
A comparison of random forest and its Gini importance
with standard chemometric methods for the feature se-
lection and classification of spectral data. BMC Bioin-
formatics, 10, 213.
[53] Gao, D., Zhang, Y.X. and Zhao, Y.H. (2009) Random
forest algorithm for classification of multi-wa velen gth data.
Research in Astronomy and Astrophysics, 9(2), 220-226.
[54] Hu, W. (2009) Identifying predictive markers of chemo-
sensitivity of breast cancer with random forests. Journal
of Biomedical Science and Engineering, 3(1), 59-64.
[55] KováccaronOVá, A., Ruttkay-Nedecký, G., Karol Haver-
líK1, I. and Janecccaronek, S. (2002) Sequence similari-
ties and evolutionary relationships of influenza virus A
hemagglutinins. Virus Genes, 24, 57-63.
[56] Colman, P.M., Hoyne, P.A. and Lawrence, M.C. (1993)
Sequence and structure alignment of paramyxovirus he-
magglutinin-neuraminidase with influenza virus neura-
minidase. Journal of Virology, 67, 2972-2980.
[57] Maurer-Stroh, S. Ma, J.M., Lee, R.T.C., Sirota, F.L. and
Eisenhaber, F. (2009) Mapping the sequence mutations
of the 2009 H1N1 influenza A virus neuraminidase rela-
tive to drug and antibody binding sites. Biology Direct, 4,
18.
[58] Liu, T. and Ye, Z.P. (2005) Attenuating mutations of the
matrix gene of influenza A/WSN/33 Virus. Journal of
Virology, 79(3), 1918-1923.
[59] Baudin, F., Petit, I., Weissenhorn, W. and Ruigrok, R.W.H.
(2001) In vitro dissection of the membrane binding and
RNP binding activities of influenza virus M1 protein.
Virology, 281, 102-108.
[60] Dua, Q.S., Wang, S.Q., Huang, R.B. and Chou, K.C.
(2010) Computational 3D structures of drug-targeting
proteins in the 2009-H1N1 influenza A virus. Chemical
Physics Letters, 485, 191-195.
[61] Ye Q., Krug R.M. and Tao Y.J. (2006) The mechanism
by which influenza A virus nucleoprotein forms oli-
gomers and binds RNA. Nature, 444, 1078-1082.
[62] Biswas, S.K., Boutz, P.L. and Nayak, D.P. (1998) Influ-
enza virus nucleoprotein interacts with influenza virus
polymerase proteins. Journal of Virology, 72, 5493-5501.
[63] Lin, D., Lan, J. and Zhang, Z. (2007) Structure and func-
tion of the NS1 protein of influenza A virus. Acta Bio-
chim Biophys Sin (Shanghai), 39(3), 155-162.
[64] Robb, N.C., Smith, M., Vreede, F.T. and Fodor, E. (2009)
NS2/NEP protein regulates transcription and replication
of the influenza virus RNA genome. Journal of General
Virology, 90, 1398-1407.
[65] Iwatsuki-Horimoto, K., Horimoto, T., Fujii, Y. and Kawa-
oka, Y. (2004) Generation of influenza A virus NS2
(NEP) mutants with an altered nuclear export signal se-
quence. Journal of Virology, 78(18), 10149-10155.
[66] Yuan, P.W., Bartlam, M., Lou, Z.Y., Chen, S.D., Zhou,
J., He, X.J., Lv, Z.Y., Ge, R.W., Li, X.M., Deng, T., Fo-
dor, E., Rao, Z.H. and Liu, Y.F. (2009) Crystal structure
of an avian influenza polymerase PAN reveals an en-
donuclease active site. Nature, 458, 909-913.
[67] Biswas, S.K. and Nayak, D.P. (1994) Mutational analysis
of the conserved motifs of influenza A virus polymerase
basic protein 1. Journal of Virology, 68, 1819-1826.
[68] Ohtsu, Y., Honda, Y., Sakata, Y., Kato, H. and Toyoda,
T. (2002) Fine mapping of the subunit binding sites of
influenza virus RNA polymerase. Microbiology and Im-
munology, 46, 167-175.
[69] Taubenberger, J.K., Reid, A.H., Lourens, R.M., Wang,
R., Jin, G. and Fanning, T.G. (2005) Characterization of
the 1918 influenza virus polymerase genes. Nature,
437(7060), 889-893.