Limits...
hiHMM: Bayesian non-parametric joint inference of chromatin state maps.

Sohn KA, Ho JW, Djordjevic D, Jeong HH, Park PJ, Kim JH - Bioinformatics (2015)

Bottom Line: To perform a principled comparison of evolutionarily distant epigenomes, we must consider species-specific biases such as differences in genome size, strength of signal enrichment and co-occurrence patterns of histone modifications.This flexible framework provides a new way to learn a consistent definition of chromatin states across multiple genomes, thus facilitating a direct comparison among them.The hierarchical and Bayesian non-parametric formulation in our approach is an important extension to the current set of methodologies for comparative chromatin landscape analysis.

View Article: PubMed Central - PubMed

Affiliation: Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department o

Show MeSH

Related in: MedlinePlus

An illustrative example using a toy simulated dataset. The heatmaps show the emission parameters of the ground truth with K = 10 chromatin states in three species (top panel) and the hiHMM and HMM estimated parameters. As State 5 (green box) and State 9 (cyan box) have different combinations of enriched marks between species, the number of distinct chromatin states across all the species is 12. Model 1 recovers the correct number of states and the enriched marks. Model 2 recovered one of the two species-specific state but missed the species-1-specific State 5. The standard HMMs miss three states even when a large K is assumed
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4481846&req=5

btv117-F2: An illustrative example using a toy simulated dataset. The heatmaps show the emission parameters of the ground truth with K = 10 chromatin states in three species (top panel) and the hiHMM and HMM estimated parameters. As State 5 (green box) and State 9 (cyan box) have different combinations of enriched marks between species, the number of distinct chromatin states across all the species is 12. Model 1 recovers the correct number of states and the enriched marks. Model 2 recovered one of the two species-specific state but missed the species-1-specific State 5. The standard HMMs miss three states even when a large K is assumed

Mentions: We compare the true simulated emission matrices and the estimated ones by each algorithm on an example dataset with K = 10 chromatin states model (Fig. 2). The true model contains two species-specific states (State 5 and State 9) that have species-specific mark combinations. Therefore, the actual number of distinct chromatin states across all the species can be viewed as 12. The genome size T was 2000, 5000 and 10 000 for each of the three species.Fig. 2.


hiHMM: Bayesian non-parametric joint inference of chromatin state maps.

Sohn KA, Ho JW, Djordjevic D, Jeong HH, Park PJ, Kim JH - Bioinformatics (2015)

An illustrative example using a toy simulated dataset. The heatmaps show the emission parameters of the ground truth with K = 10 chromatin states in three species (top panel) and the hiHMM and HMM estimated parameters. As State 5 (green box) and State 9 (cyan box) have different combinations of enriched marks between species, the number of distinct chromatin states across all the species is 12. Model 1 recovers the correct number of states and the enriched marks. Model 2 recovered one of the two species-specific state but missed the species-1-specific State 5. The standard HMMs miss three states even when a large K is assumed
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4481846&req=5

btv117-F2: An illustrative example using a toy simulated dataset. The heatmaps show the emission parameters of the ground truth with K = 10 chromatin states in three species (top panel) and the hiHMM and HMM estimated parameters. As State 5 (green box) and State 9 (cyan box) have different combinations of enriched marks between species, the number of distinct chromatin states across all the species is 12. Model 1 recovers the correct number of states and the enriched marks. Model 2 recovered one of the two species-specific state but missed the species-1-specific State 5. The standard HMMs miss three states even when a large K is assumed
Mentions: We compare the true simulated emission matrices and the estimated ones by each algorithm on an example dataset with K = 10 chromatin states model (Fig. 2). The true model contains two species-specific states (State 5 and State 9) that have species-specific mark combinations. Therefore, the actual number of distinct chromatin states across all the species can be viewed as 12. The genome size T was 2000, 5000 and 10 000 for each of the three species.Fig. 2.

Bottom Line: To perform a principled comparison of evolutionarily distant epigenomes, we must consider species-specific biases such as differences in genome size, strength of signal enrichment and co-occurrence patterns of histone modifications.This flexible framework provides a new way to learn a consistent definition of chromatin states across multiple genomes, thus facilitating a direct comparison among them.The hierarchical and Bayesian non-parametric formulation in our approach is an important extension to the current set of methodologies for comparative chromatin landscape analysis.

View Article: PubMed Central - PubMed

Affiliation: Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115, USA Department of Information and Computer Engineering, Ajou University, Suwon 443-749, South Korea, Seoul National University Biomedical Informatics (SNUBI), Division of Biomedical Informatics, Seoul National University College of Medicine, Seoul 110799, Korea, Systems Biomedical Informatics Research Center, Seoul National University, Seoul 110799, Korea, Victor Chang Cardiac Research Institute, Sydney, NSW 2010, Australia, The University of New South Wales, Sydney, NSW 2052, Australia, Center for Biomedical Informatics, Harvard Medical School, Boston, MA 02115, USA and Division of Genetics, Department o

Show MeSH
Related in: MedlinePlus