Limits...
Revealing cell cycle control by combining model-based detection of periodic expression with novel cis-regulatory descriptors.

Andersson CR, Hvidsten TR, Isaksson A, Gustafsson MG, Komorowski J - BMC Syst Biol (2007)

Bottom Line: We also find evidence for additive regulation in that the combinations of cis-regulatory descriptors associated with genes periodically expressed in fewer conditions are frequently subsets of combinations associated with genes periodically expression in more conditions.The results illustrate how a model-based approach to expression analysis may be particularly well suited to detect biologically relevant mechanisms.Our new approach makes it possible to provide more refined hypotheses about regulatory mechanisms of the cell cycle and it can easily be adjusted to reveal regulation of other, non-periodic, cellular processes.

View Article: PubMed Central - HTML - PubMed

Affiliation: The Linnaeus Centre for Bioinformatics, Uppsala University and Swedish University of Agricultural Sciences, Uppsala, Sweden. claes.andersson@lcb.uu.se

ABSTRACT

Background: We address the issue of explaining the presence or absence of phase-specific transcription in budding yeast cultures under different conditions. To this end we use a model-based detector of gene expression periodicity to divide genes into classes depending on their behavior in experiments using different synchronization methods. While computational inference of gene regulatory circuits typically relies on expression similarity (clustering) in order to find classes of potentially co-regulated genes, this method instead takes advantage of known time profile signatures related to the studied process.

Results: We explain the regulatory mechanisms of the inferred periodic classes with cis-regulatory descriptors that combine upstream sequence motifs with experimentally determined binding of transcription factors. By systematic statistical analysis we show that periodic classes are best explained by combinations of descriptors rather than single descriptors, and that different combinations correspond to periodic expression in different classes. We also find evidence for additive regulation in that the combinations of cis-regulatory descriptors associated with genes periodically expressed in fewer conditions are frequently subsets of combinations associated with genes periodically expression in more conditions. Finally, we demonstrate that our approach retrieves combinations that are more specific towards known cell-cycle related regulators than the frequently used clustering approach.

Conclusion: The results illustrate how a model-based approach to expression analysis may be particularly well suited to detect biologically relevant mechanisms. Our new approach makes it possible to provide more refined hypotheses about regulatory mechanisms of the cell cycle and it can easily be adjusted to reveal regulation of other, non-periodic, cellular processes.

Show MeSH

Related in: MedlinePlus

Method overview. We used genome-wide data including (a) binding of transcription factors in ChIP on chip experiments [16], (b) annotations of computationally inferred upstream sequence motifs [22] and (c) gene expression time profiles for different synchronization methods [2]. (d) The cis-regulatory information in a and b is combined by finding statistically associated combinations of transcription factors and motifs (i.e. cis-regulatory descriptors). (e) Periodic expression in each experiment is inferred from the temporal expression profiles. (f) Finally, machine learning is applied to model periodic expression. The model consists of rules associating cis-regulatory descriptors with patterns of periodic expression.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC2200664&req=5

Figure 1: Method overview. We used genome-wide data including (a) binding of transcription factors in ChIP on chip experiments [16], (b) annotations of computationally inferred upstream sequence motifs [22] and (c) gene expression time profiles for different synchronization methods [2]. (d) The cis-regulatory information in a and b is combined by finding statistically associated combinations of transcription factors and motifs (i.e. cis-regulatory descriptors). (e) Periodic expression in each experiment is inferred from the temporal expression profiles. (f) Finally, machine learning is applied to model periodic expression. The model consists of rules associating cis-regulatory descriptors with patterns of periodic expression.

Mentions: Although the previously reported clustering approaches generate interesting hypotheses on regulatory mechanisms, none of them make efficient use of any prior knowledge about the studied cellular process. This is in contrast to the model-based classification of expression profiles that we present here. Rather than using clustering, we employ a previously published Bayesian detector where a sinusoidal function and prior knowledge of cell division times is used to model periodic temporal profiles [19]. The detector is used to classify the temporal expression profiles as periodic or aperiodic for each of the three experiments using different synchronization methods published by Spellman et al. [2]. We argue that this class division is more suited for investigations of cell cycle regulation than a class division based on clustering (computed from, for example, Euclidean distance or correlation between expression profiles). With periodic expression as the criterion for class inclusion, the classes are directly associated with phase-specific regulation and, consequently, with the cell cycle machinery. When clustering is used, relationships between the classes and cellular processes may only be inferred from secondary data such as functional annotations. Furthermore, we show that the new approach yields understanding of regulation in terms of novel cis-regulatory descriptors. Each cis-regulatory descriptor is a binary variable that corresponds to the simultaneous presence or absence of an upstream sequence motif and observed binding of a transcription factor. Several of these interactions are supported in the literature. We then use a previously published rule-based method [7,8] to model the information available about the regulation of the periodic genes defined by the Bayesian detector as logical rules associating minimal combinations of descriptors with one or more of the periodic classes. An overview of the method is given in Figure 1.


Revealing cell cycle control by combining model-based detection of periodic expression with novel cis-regulatory descriptors.

Andersson CR, Hvidsten TR, Isaksson A, Gustafsson MG, Komorowski J - BMC Syst Biol (2007)

Method overview. We used genome-wide data including (a) binding of transcription factors in ChIP on chip experiments [16], (b) annotations of computationally inferred upstream sequence motifs [22] and (c) gene expression time profiles for different synchronization methods [2]. (d) The cis-regulatory information in a and b is combined by finding statistically associated combinations of transcription factors and motifs (i.e. cis-regulatory descriptors). (e) Periodic expression in each experiment is inferred from the temporal expression profiles. (f) Finally, machine learning is applied to model periodic expression. The model consists of rules associating cis-regulatory descriptors with patterns of periodic expression.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC2200664&req=5

Figure 1: Method overview. We used genome-wide data including (a) binding of transcription factors in ChIP on chip experiments [16], (b) annotations of computationally inferred upstream sequence motifs [22] and (c) gene expression time profiles for different synchronization methods [2]. (d) The cis-regulatory information in a and b is combined by finding statistically associated combinations of transcription factors and motifs (i.e. cis-regulatory descriptors). (e) Periodic expression in each experiment is inferred from the temporal expression profiles. (f) Finally, machine learning is applied to model periodic expression. The model consists of rules associating cis-regulatory descriptors with patterns of periodic expression.
Mentions: Although the previously reported clustering approaches generate interesting hypotheses on regulatory mechanisms, none of them make efficient use of any prior knowledge about the studied cellular process. This is in contrast to the model-based classification of expression profiles that we present here. Rather than using clustering, we employ a previously published Bayesian detector where a sinusoidal function and prior knowledge of cell division times is used to model periodic temporal profiles [19]. The detector is used to classify the temporal expression profiles as periodic or aperiodic for each of the three experiments using different synchronization methods published by Spellman et al. [2]. We argue that this class division is more suited for investigations of cell cycle regulation than a class division based on clustering (computed from, for example, Euclidean distance or correlation between expression profiles). With periodic expression as the criterion for class inclusion, the classes are directly associated with phase-specific regulation and, consequently, with the cell cycle machinery. When clustering is used, relationships between the classes and cellular processes may only be inferred from secondary data such as functional annotations. Furthermore, we show that the new approach yields understanding of regulation in terms of novel cis-regulatory descriptors. Each cis-regulatory descriptor is a binary variable that corresponds to the simultaneous presence or absence of an upstream sequence motif and observed binding of a transcription factor. Several of these interactions are supported in the literature. We then use a previously published rule-based method [7,8] to model the information available about the regulation of the periodic genes defined by the Bayesian detector as logical rules associating minimal combinations of descriptors with one or more of the periodic classes. An overview of the method is given in Figure 1.

Bottom Line: We also find evidence for additive regulation in that the combinations of cis-regulatory descriptors associated with genes periodically expressed in fewer conditions are frequently subsets of combinations associated with genes periodically expression in more conditions.The results illustrate how a model-based approach to expression analysis may be particularly well suited to detect biologically relevant mechanisms.Our new approach makes it possible to provide more refined hypotheses about regulatory mechanisms of the cell cycle and it can easily be adjusted to reveal regulation of other, non-periodic, cellular processes.

View Article: PubMed Central - HTML - PubMed

Affiliation: The Linnaeus Centre for Bioinformatics, Uppsala University and Swedish University of Agricultural Sciences, Uppsala, Sweden. claes.andersson@lcb.uu.se

ABSTRACT

Background: We address the issue of explaining the presence or absence of phase-specific transcription in budding yeast cultures under different conditions. To this end we use a model-based detector of gene expression periodicity to divide genes into classes depending on their behavior in experiments using different synchronization methods. While computational inference of gene regulatory circuits typically relies on expression similarity (clustering) in order to find classes of potentially co-regulated genes, this method instead takes advantage of known time profile signatures related to the studied process.

Results: We explain the regulatory mechanisms of the inferred periodic classes with cis-regulatory descriptors that combine upstream sequence motifs with experimentally determined binding of transcription factors. By systematic statistical analysis we show that periodic classes are best explained by combinations of descriptors rather than single descriptors, and that different combinations correspond to periodic expression in different classes. We also find evidence for additive regulation in that the combinations of cis-regulatory descriptors associated with genes periodically expressed in fewer conditions are frequently subsets of combinations associated with genes periodically expression in more conditions. Finally, we demonstrate that our approach retrieves combinations that are more specific towards known cell-cycle related regulators than the frequently used clustering approach.

Conclusion: The results illustrate how a model-based approach to expression analysis may be particularly well suited to detect biologically relevant mechanisms. Our new approach makes it possible to provide more refined hypotheses about regulatory mechanisms of the cell cycle and it can easily be adjusted to reveal regulation of other, non-periodic, cellular processes.

Show MeSH
Related in: MedlinePlus