Limits...
Computational analyses of transcriptomic data reveal the dynamic organization of the Escherichia coli chromosome under different conditions.

Ma Q, Yin Y, Schell MA, Zhang H, Li G, Xu Y - Nucleic Acids Res. (2013)

Bottom Line: Based on this hypothesis, we have predicted seven distinct sets of such domains along the E. coli genome for seven physiological conditions, namely exponential growth, stationary growth, anaerobiosis, heat shock, oxidative stress, nitrogen limitation and SOS responses.These predicted folding domains are highly stable statistically and are generally consistent with the experimental data of DNA binding sites of the nucleoid-associated proteins that assist the folding of these domains, as well as genome-scale protein occupancy profiles, hence supporting our proposed model.Our study established for the first time a strong link between a folded E. coli chromosomal structure and the encoded biological pathways and their activation frequencies.

View Article: PubMed Central - PubMed

Affiliation: Computational Systems Biology Laboratory, Department of Biochemistry and Molecular Biology and Institute of Bioinformatics, University of Georgia, Athens, GA 30602, USA.

ABSTRACT
The circular chromosome of Escherichia coli has been suggested to fold into a collection of sequentially consecutive domains, genes in each of which tend to be co-expressed. It has also been suggested that such domains, forming a partition of the genome, are dynamic with respect to the physiological conditions. However, little is known about which DNA segments of the E. coli genome form these domains and what determines the boundaries of these domain segments. We present a computational model here to partition the circular genome into consecutive segments, theoretically suggestive of the physically folded supercoiled domains, along with a method for predicting such domains under specified conditions. Our model is based on a hypothesis that the genome of E. coli is partitioned into a set of folding domains so that the total number of unfoldings of these domains in the folded chromosome is minimized, where a domain is unfolded when a biological pathway, consisting of genes encoded in this DNA segment, is being activated transcriptionally. Based on this hypothesis, we have predicted seven distinct sets of such domains along the E. coli genome for seven physiological conditions, namely exponential growth, stationary growth, anaerobiosis, heat shock, oxidative stress, nitrogen limitation and SOS responses. These predicted folding domains are highly stable statistically and are generally consistent with the experimental data of DNA binding sites of the nucleoid-associated proteins that assist the folding of these domains, as well as genome-scale protein occupancy profiles, hence supporting our proposed model. Our study established for the first time a strong link between a folded E. coli chromosomal structure and the encoded biological pathways and their activation frequencies.

Show MeSH

Related in: MedlinePlus

Boxplots showing stabilities of the predicted folding domains (exponential growth and heat shock) based on the selected MGC set versus a randomly selected MGC set as defined in the main text. The comparison among the other five pairs of predicted domain sets is shown in the left upper corner. Each box with lighter gray level represents the distance distribution between the domains predicted using the selected MGCs and domains predicted using half of the selected MGCs, and each box with darker gray level is defined similarly but against domains predicted based on randomly selected MGCs, where the y-axis is the distance axis. The Wilcoxon test P-values for each pair of distributions are shown in the top of boxes of each corresponding set of predicted folding domains.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3675479&req=5

gkt261-F2: Boxplots showing stabilities of the predicted folding domains (exponential growth and heat shock) based on the selected MGC set versus a randomly selected MGC set as defined in the main text. The comparison among the other five pairs of predicted domain sets is shown in the left upper corner. Each box with lighter gray level represents the distance distribution between the domains predicted using the selected MGCs and domains predicted using half of the selected MGCs, and each box with darker gray level is defined similarly but against domains predicted based on randomly selected MGCs, where the y-axis is the distance axis. The Wilcoxon test P-values for each pair of distributions are shown in the top of boxes of each corresponding set of predicted folding domains.

Mentions: We used the following procedure, along with the distance measure defined in ‘Materials and Methods’ section, to assess the prediction stability. Let C be the condition set used to predict a set of folding domains. We randomly selected 50% of the conditions from C, denoted it as S1 and let S2 contain 25% of conditions randomly selected from C and the same number of conditions randomly selected from the remaining portion of the 466 conditions after removing C. We then predicted the set of folding domains under conditions C, S1 and S2, denoting the three sets of predicted domains as P, P1 and P2, respectively. We performed such predictions 1000 times for each C corresponding each condition class given in Table 2 and calculated the distance distributions between P and P1 and between P and P2. Figure 2 shows the box plots of the two distributions for each of the seven classes of conditions, plus a randomly selected condition set C out of 466, with the same number of conditions to that of the above. We can clearly see that the distance between P and P1 is significantly smaller than that between P and P2 (all achieving Wilcoxon test P < 2.2e-9, shown in Figure 2) for all the seven condition sets, and there is virtually no difference for the random set. Hence, we can conclude that each predicted folding-domain set based on any of the seven classes of conditions is highly statistically significant compared with domains predicted based on randomly selected conditions, hence suggesting the strong biological significance of the predicted domains.Figure 2.


Computational analyses of transcriptomic data reveal the dynamic organization of the Escherichia coli chromosome under different conditions.

Ma Q, Yin Y, Schell MA, Zhang H, Li G, Xu Y - Nucleic Acids Res. (2013)

Boxplots showing stabilities of the predicted folding domains (exponential growth and heat shock) based on the selected MGC set versus a randomly selected MGC set as defined in the main text. The comparison among the other five pairs of predicted domain sets is shown in the left upper corner. Each box with lighter gray level represents the distance distribution between the domains predicted using the selected MGCs and domains predicted using half of the selected MGCs, and each box with darker gray level is defined similarly but against domains predicted based on randomly selected MGCs, where the y-axis is the distance axis. The Wilcoxon test P-values for each pair of distributions are shown in the top of boxes of each corresponding set of predicted folding domains.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3675479&req=5

gkt261-F2: Boxplots showing stabilities of the predicted folding domains (exponential growth and heat shock) based on the selected MGC set versus a randomly selected MGC set as defined in the main text. The comparison among the other five pairs of predicted domain sets is shown in the left upper corner. Each box with lighter gray level represents the distance distribution between the domains predicted using the selected MGCs and domains predicted using half of the selected MGCs, and each box with darker gray level is defined similarly but against domains predicted based on randomly selected MGCs, where the y-axis is the distance axis. The Wilcoxon test P-values for each pair of distributions are shown in the top of boxes of each corresponding set of predicted folding domains.
Mentions: We used the following procedure, along with the distance measure defined in ‘Materials and Methods’ section, to assess the prediction stability. Let C be the condition set used to predict a set of folding domains. We randomly selected 50% of the conditions from C, denoted it as S1 and let S2 contain 25% of conditions randomly selected from C and the same number of conditions randomly selected from the remaining portion of the 466 conditions after removing C. We then predicted the set of folding domains under conditions C, S1 and S2, denoting the three sets of predicted domains as P, P1 and P2, respectively. We performed such predictions 1000 times for each C corresponding each condition class given in Table 2 and calculated the distance distributions between P and P1 and between P and P2. Figure 2 shows the box plots of the two distributions for each of the seven classes of conditions, plus a randomly selected condition set C out of 466, with the same number of conditions to that of the above. We can clearly see that the distance between P and P1 is significantly smaller than that between P and P2 (all achieving Wilcoxon test P < 2.2e-9, shown in Figure 2) for all the seven condition sets, and there is virtually no difference for the random set. Hence, we can conclude that each predicted folding-domain set based on any of the seven classes of conditions is highly statistically significant compared with domains predicted based on randomly selected conditions, hence suggesting the strong biological significance of the predicted domains.Figure 2.

Bottom Line: Based on this hypothesis, we have predicted seven distinct sets of such domains along the E. coli genome for seven physiological conditions, namely exponential growth, stationary growth, anaerobiosis, heat shock, oxidative stress, nitrogen limitation and SOS responses.These predicted folding domains are highly stable statistically and are generally consistent with the experimental data of DNA binding sites of the nucleoid-associated proteins that assist the folding of these domains, as well as genome-scale protein occupancy profiles, hence supporting our proposed model.Our study established for the first time a strong link between a folded E. coli chromosomal structure and the encoded biological pathways and their activation frequencies.

View Article: PubMed Central - PubMed

Affiliation: Computational Systems Biology Laboratory, Department of Biochemistry and Molecular Biology and Institute of Bioinformatics, University of Georgia, Athens, GA 30602, USA.

ABSTRACT
The circular chromosome of Escherichia coli has been suggested to fold into a collection of sequentially consecutive domains, genes in each of which tend to be co-expressed. It has also been suggested that such domains, forming a partition of the genome, are dynamic with respect to the physiological conditions. However, little is known about which DNA segments of the E. coli genome form these domains and what determines the boundaries of these domain segments. We present a computational model here to partition the circular genome into consecutive segments, theoretically suggestive of the physically folded supercoiled domains, along with a method for predicting such domains under specified conditions. Our model is based on a hypothesis that the genome of E. coli is partitioned into a set of folding domains so that the total number of unfoldings of these domains in the folded chromosome is minimized, where a domain is unfolded when a biological pathway, consisting of genes encoded in this DNA segment, is being activated transcriptionally. Based on this hypothesis, we have predicted seven distinct sets of such domains along the E. coli genome for seven physiological conditions, namely exponential growth, stationary growth, anaerobiosis, heat shock, oxidative stress, nitrogen limitation and SOS responses. These predicted folding domains are highly stable statistically and are generally consistent with the experimental data of DNA binding sites of the nucleoid-associated proteins that assist the folding of these domains, as well as genome-scale protein occupancy profiles, hence supporting our proposed model. Our study established for the first time a strong link between a folded E. coli chromosomal structure and the encoded biological pathways and their activation frequencies.

Show MeSH
Related in: MedlinePlus