Advertisement
Table 2 Top features in CoreBoost
From: Boosting with stumps for predicting transcription start sites
| Classifier type | Features | |
|---|---|---|
| CpG | P versus U | Log-likelihood ratios from third order Markov chain, log-likelihood ratios from TSS weight matrix |
| GC-box score, weighted score of transcription factor NFY, weighted energy score at position +1 | ||
| Weighted score of transcription factor YY1, TATA score, weighted score of transcription factor ELK1 | ||
| MTE score, weighted score of transcription factor CREB | ||
| P versus D | Log-likelihood ratios from third order Markov chain, GC-box score | |
| Weighted score of transcription factor NFY | ||
| Log-likelihood ratios from TSS weight matrix | ||
| Difference between the energy score around positions -25 and +1 and the average from surroundings | ||
| Log-likelihood ratios from transcription factor ELK1, frequency of G+C | ||
| Log-likelihood ratios from transcription factor YY1, TATA score, frequency of G | ||
| Non-CpG | P versus U | Correlation between vector of energy scores and empirical average energy profile |
| Log-likelihood ratios from third order Markov chain, TATA score | ||
| Difference between the energy score around positions -25 and +1 and the average from surroundings | ||
| Weighted energy at position +1 | ||
| Proportion of Inr and GC-box pair within 10 bp of observed distance, Inr score. | ||
| P versus D | Correlation between vector of energy scores and empirical average energy profile, TATA score | |
| Log-likelihood ratios from third order Markov chain | ||
| Weighted energy at position +1 | ||
| Correlation between vector of flexibility scores and empirical average flexibility profile, Inr score | ||
| Difference between the flexibility score around position +1 and the average from surroundings, GC-box score | ||
