Median Pearson correlation coefficient between Hi-C scores and the prediction scores from epiphany, C. Origami, DynamicEdgeConv, GAT, GCN, and Chrombus. All models were trained and tested on the GM12878 cell line. Chrombus outperforms the other methods in predicting interactions at distances of 1–2 Mb and beyond 2 Mb.
| Methods . | 0–1 M . | 1–2 M . | >2 M . | Overall . |
|---|---|---|---|---|
| Epiphany | 0.834 | – | – | 0.834 |
| C.Origami | 0.795 | 0.483 | – | 0.798 |
| DynamicEdgeConv | 0.294 | 0.252 | 0.442 | 0.728 |
| GAT | 0.050 | −0.014 | 0.000 | 0.050 |
| GCN | 0.019 | 0.028 | 0.039 | 0.050 |
| Chrombus | 0.755 | 0.540 | 0.582 | 0.908 |
| Methods . | 0–1 M . | 1–2 M . | >2 M . | Overall . |
|---|---|---|---|---|
| Epiphany | 0.834 | – | – | 0.834 |
| C.Origami | 0.795 | 0.483 | – | 0.798 |
| DynamicEdgeConv | 0.294 | 0.252 | 0.442 | 0.728 |
| GAT | 0.050 | −0.014 | 0.000 | 0.050 |
| GCN | 0.019 | 0.028 | 0.039 | 0.050 |
| Chrombus | 0.755 | 0.540 | 0.582 | 0.908 |
The bold values indicate the best predictive performance obtained from the corresponding methods and the dash symbols mean the corresponding methods are not applicable in those ranges.
Median Pearson correlation coefficient between Hi-C scores and the prediction scores from epiphany, C. Origami, DynamicEdgeConv, GAT, GCN, and Chrombus. All models were trained and tested on the GM12878 cell line. Chrombus outperforms the other methods in predicting interactions at distances of 1–2 Mb and beyond 2 Mb.
| Methods . | 0–1 M . | 1–2 M . | >2 M . | Overall . |
|---|---|---|---|---|
| Epiphany | 0.834 | – | – | 0.834 |
| C.Origami | 0.795 | 0.483 | – | 0.798 |
| DynamicEdgeConv | 0.294 | 0.252 | 0.442 | 0.728 |
| GAT | 0.050 | −0.014 | 0.000 | 0.050 |
| GCN | 0.019 | 0.028 | 0.039 | 0.050 |
| Chrombus | 0.755 | 0.540 | 0.582 | 0.908 |
| Methods . | 0–1 M . | 1–2 M . | >2 M . | Overall . |
|---|---|---|---|---|
| Epiphany | 0.834 | – | – | 0.834 |
| C.Origami | 0.795 | 0.483 | – | 0.798 |
| DynamicEdgeConv | 0.294 | 0.252 | 0.442 | 0.728 |
| GAT | 0.050 | −0.014 | 0.000 | 0.050 |
| GCN | 0.019 | 0.028 | 0.039 | 0.050 |
| Chrombus | 0.755 | 0.540 | 0.582 | 0.908 |
The bold values indicate the best predictive performance obtained from the corresponding methods and the dash symbols mean the corresponding methods are not applicable in those ranges.
This PDF is available to Subscribers Only
View Article Abstract & Purchase OptionsFor full access to this pdf, sign in to an existing account, or purchase an annual subscription.
