Identification of human-to-human transmissibility factors in PB2 proteins of influenza A by large-scale mutual information analysis

Miotto, Olivo; Heiny, A. T.; Tan, Tin Wee; August, J. Thomas; Brusic, Vladimir

doi:10.1186/1471-2105-9-s1-s18

Cited by 99 publications

(116 citation statements)

References 45 publications

Supporting

Mentioning

115

Contrasting

Order By: Relevance

“…It selects a set of positions whose combination provides optimal discrimination between two groups (e.g., pH1N1 and swine-H1N1 sequences). For comparison we analyzed the same datasets using a mutual-information-based method (29). Reassuringly, the majority of the highly ranked positions presented here were also obtained in that analysis (SI Text S3 and Tables S5 and S6).…”

Section: Discussionmentioning

confidence: 99%

“…As manual inspection of this volume of data is virtually impossible; it is necessary to use computational methods. Entropy (mutual-information)-based methods that use the amino acid frequency in each position are commonly used to this effect (27)(28)(29). These methods consider each position in the alignment separately, disregarding the possible relations between the sequence positions.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Putative amino acid determinants of the emergence of the 2009 influenza A (H1N1) virus in the human population

Meroz

Yoon

Ducatez

et al. 2011

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

The emergence of the unique H1N1 influenza A virus in 2009 resulted in a pandemic that has spread to over 200 countries. The constellation of molecular factors leading to the emergence of this strain is still unclear. Using a computational approach, we identified molecular determinants that may discriminate the hemagglutinin protein of the 2009 human pandemic H1N1 (pH1N1) strain from that of other H1N1 strains. As expected, positions discriminating the pH1N1 from seasonal human strains were located in or near known H1N1 antigenic sites, thus camouflaging the pH1N1 strain from immune recognition. For example, the alteration S145K (an antigenic position) was found as a characteristic of the pH1N1 strain. We also detected positions in the hemagglutinin protein differentiating classical swine viruses from pH1N1. These positions were mostly located in and around the receptor-binding pocket, possibly influencing binding affinity to the human cell. Such alterations may be liable in part for the virus's efficient infection and adaptation to humans. For instance, 133 A and 149 were identified as discriminative positions. Significantly, we showed that the substitutions R133 A K and R149K, predicted to be pH1N1 characteristics, each altered virus binding to erythrocytes and conferred virulence to A/swine/NC/18161/02 in mice, reinforcing the computational findings. Our findings provide a structural explanation for the deficient immunity of humans to the pH1N1 strain. Moreover, our analysis points to unique molecular factors that may have facilitated the emergence of this swine variant in humans, in contrast to other swine variants that failed.A pandemic H1N1 (pH1N1) human influenza virus was identified in April 2009 (1) and has since spread to over 200 countries and caused over 18,000 deaths (2). Evolutionary analysis of the pH1N1 strain indicates that its HA belongs to the classical swine lineage. HAs in this lineage are more similar to those of historical human strains such as the 1918 H1N1 strain than to those of circulating H1N1 strains from recent years (3). Prior to the emergence of the pH1N1 strain, only sporadic cases of human infection by swine influenza viruses had been reported (4-8). The molecular basis enabling this recent strain to efficiently infect and be transmitted between humans remains obscure (9-12).Herein, we used a computational approach to uncover signature residues in the receptor-binding domain (RBD) of the pH1N1 HA protein. Using a machine-learning algorithm of alternating decision trees (ADTs) (13), we predicted positions that differentiated between HA protein sequences of the pH1N1 strain and HA sequences of other H1N1 strains. The algorithm combines moderately successful rules to produce highly accurate predictions (13). Here the rules represent correlations between the presence of specific amino acid type(s) in selected positions and the sequence annotation, e.g., pandemic versus seasonal strain. This is done by iteratively reweighting training examples (i.e., HA sequences), thus, concentratin...

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

Putative amino acid determinants of the emergence of the 2009 influenza A (H1N1) virus in the human population

Meroz

Yoon

Ducatez

et al. 2011

Proc. Natl. Acad. Sci. U.S.A.

View full text Add to dashboard Cite

show abstract

“…In our earlier report, the consensus approach has been used for epitope prediction of H1N1 HA and NA proteins (28). The current study also employed different prediction tools: NetCTL 1.2, SYFPEITHI, and BI-MAS (26,34,39) for class I binding (CD8+ T cell) epitopes and Immune Epitope Database (IEDB)-based SMM align, ProPred and NetMHC II for class II binding (CD4+ T cell) epitopes (32,33,43).…”

Section: Identification Of T Cell Epitopesmentioning

confidence: 99%

“…MUSCLE and AVANA tools were used to obtain the conserved peptide stretches ( ‡90%) (17,27,28). Conservation refers to identical peptide sequences that are present in at least 90% of total M1 protein sequences, and these conserved peptide sequences were used for further epitope prediction.…”

Section: Sequence Retrieval and Conservation Analysismentioning

confidence: 99%

Identification of Conserved Peptides Comprising Multiple T Cell Epitopes of Matrix 1 Protein in H1N1 Influenza Virus

Lohia

Baranwal

2015

Viral Immunology

View full text Add to dashboard Cite

Cell mediated immune response plays a key role in combating viral infection and thus identification of new vaccine targets manifesting T cell mediated response may serve as an ideal approach for influenza vaccine. The present study involves the application of an immunoinformatics-based consensus approach for epitope prediction (three epitope prediction tools each for CD4+ and CD8+ T cell epitopes) and molecular docking to identify peptide sequences containing T cell epitopes using the conserved sequences from all the Matrix 1 protein sequences of H1N1 virus available until April 2015. Three peptides comprising CD4+ and CD8+ T cell epitopes were obtained, which were not exactly reported in earlier studies. Population coverage study of these multiepitope peptides revealed that they are capable of inducing a potent immune response belonging to individuals from different populations and ethnicity distributed around the globe. Conservation study with other subtypes of influenza virus infecting humans (H2N2, H5N1, H7N9, and H3N2) revealed that these three peptides were conserved (>90%), with 100% identity in most of these strains. Hence, these peptides can impart immunity against H1N1 as well as other subtypes of influenza virus. A molecular docking study of the predicted peptides with class I and II human leukocyte antigen (HLA) molecules has shown that the majority of them have comparable binding energies to that of native peptides. Hence, these peptides from Matrix 1 protein of H1N1 appear to be promising candidates for universal vaccine design.

show abstract

“…Support vector machines, entropy, and mutual information were utilized to uncover the unique molecular features of these viruses [14][15][16][17][18][19][20][21][22]. Most recently, Random Forests [23] were applied successfully to tackle the same problem, where novel host markers in the proteins and genes of pandemic 2009 H1N1 were identified [24,25].…”

Section: Introductionmentioning

confidence: 99%

Characterization of asian and north American avian H5N1

2011

AJMB

View full text Add to dashboard Cite

Since the emerge of the highly pathogenic avian H5N1 virus in Asia in 1996, the possibility for this virus to cross species barriers to infect humans and its ability to cause large outbreaks in birds have been a public health concern. This virus has been spreading from Asia to Europe and Africa by migratory birds with North America as its next possible stop. In this study, an ensemble of computational techniques including Random Forests, Informational Spectrum Method, Entropy, and Mutual Information were employed to unravel the distinct characteristics of Asian and North American avian H5N1 in comparison with human and swine H5N1. Critical differences were identified in the HA cleavage and binding sites, the HA receptor selection, the interaction patterns of HA and NA, and NP, PA, PB1, and PB2, and the important sites in the influenza proteins including HA, NA, M1, M2, NS1, NS2, NP, PA, PB1, PB1-F2, and PB2.

show abstract

Identification of human-to-human transmissibility factors in PB2 proteins of influenza A by large-scale mutual information analysis

Cited by 99 publications

References 45 publications

Putative amino acid determinants of the emergence of the 2009 influenza A (H1N1) virus in the human population

Putative amino acid determinants of the emergence of the 2009 influenza A (H1N1) virus in the human population

Identification of Conserved Peptides Comprising Multiple T Cell Epitopes of Matrix 1 Protein in H1N1 Influenza Virus

Characterization of asian and north American avian H5N1

Contact Info

Product

Resources

About