The structure of chromosomes is largely dependent on primary DNA sequence though the interaction is not fully understood yet. Hilbert-CNN developed by Yin et. al. takes the primary DNA sequence as input and predicts key determinants of the chromatin structure. The usage of large convolution filter at an early stage in the network facilitates the detection of long-term interactions. Hilbert curves (space-filling curves) are used to map DNA sequences to higher dimensional images. Every element of the sequence is mapped to a pixel in 2D image. This causes the proximal elements to stay close and reduces the distance between distal elements in the DNA sequence. This image representation is fed as input to a convolutional neural network which predicts the chromatin state of the DNA sequence.
Input variables : Primary DNA sequence
Output Variables : Predicted chromatin state
| Statistical | : | Somers D | Accuracy | Precision and Recall | Confusion Matrix | F1 Score | Roc and Auc | Prevalence | Detection Rate | Balanced Accuracy | Cohen's Kappa | Concordance | Gini Coefficent | KS Statistic | Youden's J Index |
| Infrastructure | : | Log Bytes | Logging/User/IAMPolicy | Logging/User/VPN | CPU Utilization | Memory Usage | Error Count | Prediction Count | Prediction Latencies | Private Endpoint Prediction Latencies | Private Endpoint Response Count |
Visit Model : github.com
Additional links : arxiv.org
| Model Category | : | Public |
| Date Published | : | June, 2018 |
| Healthcare Domain | : | Life Sciences |
| Code | : | github.com |
| Clinical Information Extraction |
| Insight Extraction |