Truncation utility module
ScoreVariances¶Bases: tuple
| Attributes: |
|---|
Methods
count(…) |
|
index((value, [start, …) |
Raises ValueError if the value is not present. |
idx¶Alias for field number 0
resSeq¶Alias for field number 1
variance¶Alias for field number 2
Truncation[source]¶Bases: object
Holds information relating to a single truncation of a cluster of models
| Attributes: |
|
|---|
num_residues¶Truncator(work_dir)[source]¶Bases: object
Methods
calculate_truncations([models, …]) |
Returns a list of Truncation objects, one for each truncation level. |
truncate_models(models[, max_cluster_size, …]) |
Generate a set of Truncation objects, referencing a set of truncated models generated from the supplied models |
calculate_truncations(models=None, truncation_method=None, percent_truncation=None, truncation_pruning=None, residue_scores=None, alignment_file=None, homologs=False)[source]¶Returns a list of Truncation objects, one for each truncation level.
This method doesn’t do any truncating - it just calculates the data for each truncation level.
truncate_models(models, max_cluster_size=200, truncation_method=None, percent_truncation=None, truncation_pruning=None, residue_scores=None, homologs=False, alignment_file=None, work_dir=None)[source]¶Generate a set of Truncation objects, referencing a set of truncated models generated from the supplied models
calculate_residues_focussed(var_by_res)[source]¶The sweet spot for success seems to occur in the interval 5-40 residues. Up till now we have always worked in 5% intervals, so 20 truncation levels The new strategy is to ensure that always have at least half of the truncations in the interval < 40 residues => 10 truncations in 40, so at least 4 residue chunks in this interval.
The strategy is therefore for < 80 residues, just split evenly into 20 chunks.
For > 80 residues, split < 40 into 10 4-residue chunks, and split the interval 40 -> end into 10 even chunks.
calculate_residues_percent(var_by_res, percent_interval)[source]¶Calculate the list of residues to keep if we are keeping self.percent residues under each truncation bin. The threshold is just the threshold of the most variable residue