Search

Additional file 2. Review history.

Additional file 1. Supplementary text, methods, figures, and tables.

Additional file 1. Supplementary materials (Supplementary Figures S1���S4, Supplementary Tables S1).

Additional file 19: Table S15. Examples of rules generated by OneR, JRip and PART for two-class and three-class H5N1 datasets containing concatenated alignments of IAV proteins.

Additional file 11: Table S7. Extrapolated incomplete IAV genomes.

Additional file 16: Table S12. Examples of rules generated by OneR, JRip and PART for two-class and three-class C57BL/6 datasets containing concatenated alignments of IAV proteins.

Additional file 17: Table S13. Examples of rules generated by OneR, JRip and PART for two-class and three-class H1N1 datasets containing concatenated alignments of IAV proteins.

Additional file 13: Table S9. Examples of rules generated by OneR, JRip and PART for two-class and three-class MIV datasets containing concatenated alignments of IAV proteins.

Additional file 4; Figure S4. Multiple comparisons between mean accuracies of OneR, JRip and PART models for IAV virulence based on two-class and three-class BALB/C, C57BL/6, H1N1, H3N2 and H5N1 datasets containing either the concatenated alignment of all IAV proteins or an individual alignment of PB2, PB1, PA, HA, NP, NA, M1, NS1, PB1-F2, PA-X, M2 and NS2 proteins.

Additional file 3: Figure S3. Multiple comparisons between mean accuracies of OneR, JRip and PART models for IAV virulence based on two-class and three-class MIV and IV datasets containing either the concatenated alignment of all IAV proteins or an individual alignment of PB2, PB1, PA, HA, NP, NA, M1, NS1, PB1-F2, PA-X, M2 and NS2 proteins.

Additional file 12: Table S8. Extrapolated partial IAV segments.

Additional file 1. Supplementary materials (Supplementary Tables S1–S3, Supplementary Figures S1–S10).

Additional file 7: Table S3. Reduction of multiple records for infection of a specific IAV strain in different mouse strains into a single record (with supplementary references).

Additional file 18: Table S14. Examples of rules generated by OneR, JRip and PART for two-class and three-class H3N2 datasets containing concatenated alignments of IAV proteins.

Additional file 8: Table S4. Metadata of IAV nucleotide sequences used in this study (with supplementary references).

Additional file 5: Table S1. Initial dataset for IAV infections in mice with virulence information (with supplementary references).

Additional file 2: Figure S2. Accuracy distribution of 100 PART/random forest models learned independently from two-class and three-class datasets containing the concatenated alignments of IAV proteins.

Additional file 9: Table S5. Mutant and reassortant IAVs generated in this study.

Additional file 14: Table S10. Examples of rules generated by OneR, JRip and PART for two-class and three-class IV datasets containing concatenated alignments of IAV proteins.

Additional file 10: Table S6. GISAID acknowledgement table for sequences used in this study.

Additional file 15: Table S11. Examples of rules generated by OneR, JRip and PART for two-class and three-class BALB/C datasets containing concatenated alignments of IAV proteins.

Additional file 6: Table S2. Reduction of multiple records for infection involving specific IAV and mouse strains into a single record (with supplementary references).

Additional file 1: Figure S1. Accuracy distribution of 100 OneR/JRip/PART models learned independently from two-class and three-class BALB/C, C57BL/6, H1N1, H3N2, and H5N1 datasets containing either the concatenated alignments of all IAV proteins or an individual alignment of PB2, PB1, PA, HA, NP, NA, M1, NS1, PB1-F2, PA-X, M2 or NS2 proteins.