Skip to content

Results and Interpretation

Jump to bottom
Julius Wiegert edited this page Dec 18, 2023 · 3 revisions

Overview

EBG stores its results in a folder located in the current directory with the name -o OUTPUT_NAME. If you run in mode -t b it will provide the direct bootstrap value predictions, lower bounds of them, and the probabilities, that the branches exceed a threshold of 70, 75, 80, or 85 with their prediction uncertainty. In this mode, EBG will output a total of 11 trees. Each tree contains its predictions in the Newick-format field for support.
The results in the result folder might look like:

  • OUTPUT_NAME_bs_over_70_support_prediction.newick
    Tree with probabilities as support values. They quantify the probability, that the branch exceeds a bootstrap support of 70.

  • OUTPUT_NAME_bs_over_75_support_prediction.newick
    Tree with probabilities as support values. They quantify the probability, that the branch exceeds a bootstrap support of 75.

  • OUTPUT_NAME_bs_over_80_support_prediction.newick
    Tree with probabilities as support values. They quantify the probability, that the branch exceeds a bootstrap support of 80.

  • OUTPUT_NAME_bs_over_85_support_prediction.newick
    Tree with probabilities as support values. They quantify the probability, that the branch exceeds a bootstrap support of 85.

  • OUTPUT_NAME_features.csv
    File that contains all the features computed by EBG

  • OUTPUT_NAME_lower5_support_prediction.newick
    Tree with the 5% lower bound support predictions.

  • OUTPUT_NAME_lower10_support_prediction.newick
    Tree with the 10% lower bound support predictions.

  • OUTPUT_NAME_median_support_prediction.newick
    Tree with the median support prediction.

  • OUTPUT_NAME_uncertainty_bs_over_70_support_prediction.newick
    Entropy-based uncertainty of the predictions in the bs_over_70_support_prediction.newick tree.

  • OUTPUT_NAME_uncertainty_bs_over_75_support_prediction.newick
    Entropy-based uncertainty of the predictions in the bs_over_75_support_prediction.newick tree.

  • OUTPUT_NAME_uncertainty_bs_over_80_support_prediction.newick
    Entropy-based uncertainty of the predictions in the bs_over_80_support_prediction.newick tree.

  • OUTPUT_NAME_uncertainty_bs_over_85_support_prediction.newick
    Entropy-based uncertainty of the predictions in the bs_over_85_support_prediction.newick tree.

  • tmp
    This folder contains all the temporary files EBG needs including the parsimony bootstraps.

    important: This folder can quickly become very large if you are using large phylogenies as EBG stores all resampled MSAs and trees. If you are only interested in the prediction results, you can delete the tmp folder.

Median and Lower Bound Predictions

The 5/10% lower bound indicates a belief that there is only a 5/10% probability that the true bootstrap value falls below this estimate. On the other hand, the median prediction serves as a point estimate, signifying that EBG expects 50% of the actual support values to be below it and 50% to be above it. If you seek a straightforward estimate of your bootstrap values, the median prediction provides the value you are looking for.

Uncertainty

If you want to decide whether a prediction in OUTPUT_NAME_uncertainty_bs_over_70_support_prediction.newick is trustworthy, you can have a look at the uncertainties in OUTPUT_NAME_uncertainty_bs_over_70_support_prediction.newick. They contain the Shannon entropy of the prediction probabilities and their counter-probability.

Clone this wiki locally