Limits...
Distinct and predictive histone lysine acetylation patterns at promoters, enhancers, and gene bodies.

Rajagopal N, Ernst J, Ray P, Wu J, Zhang M, Kellis M, Ren B - G3 (Bethesda) (2014)

Bottom Line: Unexpectedly, we found that histone acetylation alone performs well in distinguishing these unique genomic regions.Further, we found the association of characteristic acetylation patterns with genic regions and association of chromatin state with splicing.Taken together, our work underscores the diverse functional roles of histone acetylation in gene regulation and provides several testable hypotheses to dissect these roles.

View Article: PubMed Central - PubMed

Affiliation: Ludwig Institute for Cancer Research, 9500 Gilman Drive, La Jolla, California 92093-0653 Bioinformatics and Systems Biology Program, University of California, San Diego, La Jolla, California 92037 Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139.

Show MeSH
Genome-wide prediction of promoters. Receiver operating characteristic (ROC) curves for prediction of promoters in (A) 1H and (B) IMR90 using all 24 modifications (blue), H3K4me3 (black), H3K4me1/2/3 (red), or all 15 acetylations (green). Out-of-bag variable importance for acetylations in making genome-wide prediction of promoters in (C) 1H and (D) IMR90. Modification names indicated in red are the ones that show top-most variable importance in both cell types and are considered candidates for selection in the minimal set. ROC curves for prediction of promoters using various minimal combinations of acetylations in (E) 1H and (F) IMR90, as compared with the prediction using all 15 acetylations (in blue).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4232531&req=5

fig2: Genome-wide prediction of promoters. Receiver operating characteristic (ROC) curves for prediction of promoters in (A) 1H and (B) IMR90 using all 24 modifications (blue), H3K4me3 (black), H3K4me1/2/3 (red), or all 15 acetylations (green). Out-of-bag variable importance for acetylations in making genome-wide prediction of promoters in (C) 1H and (D) IMR90. Modification names indicated in red are the ones that show top-most variable importance in both cell types and are considered candidates for selection in the minimal set. ROC curves for prediction of promoters using various minimal combinations of acetylations in (E) 1H and (F) IMR90, as compared with the prediction using all 15 acetylations (in blue).

Mentions: Using all 24 histone modifications, our approach can accurately predict promoters with ∼92% true-positive (TP) rate and ∼1.6% false-positive (FP) rate in 1H, whereas in IMR90 we observed even better performance (TP ∼95%, FP ∼0.3%) (Figure 2, A and B). Using the out-of-bag variable measure, we identified H3K4me3 as the most informative mark required to predict promoters, followed by H3K4me2 and H3K4me1 (Figure S2, A and B). In terms of the area under the curve (AUC), this minimal set performs comparably with the set of all 24 modifications in both 1H and IMR90 (AUCmin/AUCall = 0.99) (Figure 2A, red vs. blue). While in 1H, this set is comparable with using just H3K4me3 (Figure 2A, black vs. red); in IMR90, the addition of the two marks leads to an improvement of ∼10% in TP rate as compared with H3K4me3 (Figure 2B, black vs. red).


Distinct and predictive histone lysine acetylation patterns at promoters, enhancers, and gene bodies.

Rajagopal N, Ernst J, Ray P, Wu J, Zhang M, Kellis M, Ren B - G3 (Bethesda) (2014)

Genome-wide prediction of promoters. Receiver operating characteristic (ROC) curves for prediction of promoters in (A) 1H and (B) IMR90 using all 24 modifications (blue), H3K4me3 (black), H3K4me1/2/3 (red), or all 15 acetylations (green). Out-of-bag variable importance for acetylations in making genome-wide prediction of promoters in (C) 1H and (D) IMR90. Modification names indicated in red are the ones that show top-most variable importance in both cell types and are considered candidates for selection in the minimal set. ROC curves for prediction of promoters using various minimal combinations of acetylations in (E) 1H and (F) IMR90, as compared with the prediction using all 15 acetylations (in blue).
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4232531&req=5

fig2: Genome-wide prediction of promoters. Receiver operating characteristic (ROC) curves for prediction of promoters in (A) 1H and (B) IMR90 using all 24 modifications (blue), H3K4me3 (black), H3K4me1/2/3 (red), or all 15 acetylations (green). Out-of-bag variable importance for acetylations in making genome-wide prediction of promoters in (C) 1H and (D) IMR90. Modification names indicated in red are the ones that show top-most variable importance in both cell types and are considered candidates for selection in the minimal set. ROC curves for prediction of promoters using various minimal combinations of acetylations in (E) 1H and (F) IMR90, as compared with the prediction using all 15 acetylations (in blue).
Mentions: Using all 24 histone modifications, our approach can accurately predict promoters with ∼92% true-positive (TP) rate and ∼1.6% false-positive (FP) rate in 1H, whereas in IMR90 we observed even better performance (TP ∼95%, FP ∼0.3%) (Figure 2, A and B). Using the out-of-bag variable measure, we identified H3K4me3 as the most informative mark required to predict promoters, followed by H3K4me2 and H3K4me1 (Figure S2, A and B). In terms of the area under the curve (AUC), this minimal set performs comparably with the set of all 24 modifications in both 1H and IMR90 (AUCmin/AUCall = 0.99) (Figure 2A, red vs. blue). While in 1H, this set is comparable with using just H3K4me3 (Figure 2A, black vs. red); in IMR90, the addition of the two marks leads to an improvement of ∼10% in TP rate as compared with H3K4me3 (Figure 2B, black vs. red).

Bottom Line: Unexpectedly, we found that histone acetylation alone performs well in distinguishing these unique genomic regions.Further, we found the association of characteristic acetylation patterns with genic regions and association of chromatin state with splicing.Taken together, our work underscores the diverse functional roles of histone acetylation in gene regulation and provides several testable hypotheses to dissect these roles.

View Article: PubMed Central - PubMed

Affiliation: Ludwig Institute for Cancer Research, 9500 Gilman Drive, La Jolla, California 92093-0653 Bioinformatics and Systems Biology Program, University of California, San Diego, La Jolla, California 92037 Computer Science and Artificial Intelligence Laboratory, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139.

Show MeSH