Limits...
EBSeq-HMM: a Bayesian approach for identifying gene-expression changes in ordered RNA-seq experiments.

Leng N, Li Y, McIntosh BE, Nguyen BK, Duffin B, Tian S, Thomson JA, Dewey CN, Stewart R, Kendziorski C - Bioinformatics (2015)

Bottom Line: With improvements in next-generation sequencing technologies and reductions in price, ordered RNA-seq experiments are becoming common.EBSeq-HMM may also be used for inference regarding isoform expression.An R package containing examples and sample datasets is available at Bioconductor. kendzior@biostat.wisc.edu Supplementary data are available at Bioinformatics online.

View Article: PubMed Central - PubMed

Affiliation: Department of Statistics, University of Wisconsin, Madison, WI, USA, Regenerative Biology, Morgridge Institute for Research, Madison, WI, USA.

No MeSH data available.


Related in: MedlinePlus

(a) An auto-regressive hidden Markov component models dynamic paths. (b) An auto-regressive non-hidden Markov component models constant and sporadic paths
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4528625&req=5

btv193-F1: (a) An auto-regressive hidden Markov component models dynamic paths. (b) An auto-regressive non-hidden Markov component models constant and sporadic paths

Mentions: In summary, the time-course for a dynamic gene is governed by two interrelated probabilistic mechanisms: the conditional distribution (emissions model) at each time and the process describing the evolution of states over time. Initially, we assume that the observed expression vector can be characterized by the Beta-NB model described earlier and that the state process can be described by a Markov chain. Were it the case that dependence among measurements is fully captured by the state process, the proposed model would be a standard HMM. However, this last assumption does not hold, given that Xt for dynamic genes depends not only on the state but also on through . Consequently, the model for dynamic genes is given by a Markov-switching auto-regressive model, as in Hamilton (1989) and Ailliot and Monbet (2012) (Fig. 1). For constant and sporadic genes, we assume the same emissions model, but do not assume the state process is Markov. Taken together, since we do not know the expression path type a priori, the model for the full set of expression measurements is a two-component mixture over the sporadic/constant and dynamic genes.Fig. 1.


EBSeq-HMM: a Bayesian approach for identifying gene-expression changes in ordered RNA-seq experiments.

Leng N, Li Y, McIntosh BE, Nguyen BK, Duffin B, Tian S, Thomson JA, Dewey CN, Stewart R, Kendziorski C - Bioinformatics (2015)

(a) An auto-regressive hidden Markov component models dynamic paths. (b) An auto-regressive non-hidden Markov component models constant and sporadic paths
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4528625&req=5

btv193-F1: (a) An auto-regressive hidden Markov component models dynamic paths. (b) An auto-regressive non-hidden Markov component models constant and sporadic paths
Mentions: In summary, the time-course for a dynamic gene is governed by two interrelated probabilistic mechanisms: the conditional distribution (emissions model) at each time and the process describing the evolution of states over time. Initially, we assume that the observed expression vector can be characterized by the Beta-NB model described earlier and that the state process can be described by a Markov chain. Were it the case that dependence among measurements is fully captured by the state process, the proposed model would be a standard HMM. However, this last assumption does not hold, given that Xt for dynamic genes depends not only on the state but also on through . Consequently, the model for dynamic genes is given by a Markov-switching auto-regressive model, as in Hamilton (1989) and Ailliot and Monbet (2012) (Fig. 1). For constant and sporadic genes, we assume the same emissions model, but do not assume the state process is Markov. Taken together, since we do not know the expression path type a priori, the model for the full set of expression measurements is a two-component mixture over the sporadic/constant and dynamic genes.Fig. 1.

Bottom Line: With improvements in next-generation sequencing technologies and reductions in price, ordered RNA-seq experiments are becoming common.EBSeq-HMM may also be used for inference regarding isoform expression.An R package containing examples and sample datasets is available at Bioconductor. kendzior@biostat.wisc.edu Supplementary data are available at Bioinformatics online.

View Article: PubMed Central - PubMed

Affiliation: Department of Statistics, University of Wisconsin, Madison, WI, USA, Regenerative Biology, Morgridge Institute for Research, Madison, WI, USA.

No MeSH data available.


Related in: MedlinePlus