calc_rank_of_mrca_bounds_between

calc_rank_of_mrca_bounds_between(first: Union[HereditaryStratigraphicColumn, HereditaryStratigraphicSpecimen], second: Union[HereditaryStratigraphicColumn, HereditaryStratigraphicSpecimen], prior: Literal['arbitrary'], confidence_level: float = 0.95, strict=True) → Tuple[int, int] | None

Within what generation range did MRCA fall?

Calculate bounds on estimate for the number of depositions elapsed along the line of descent before the most recent common ancestor with second.

Parameters

prior{“arbitrary”}

Prior probability density distribution over possible generations of the MRCA.

Currently only “arbitrary” supported.

confidence_levelfloat, optional

Bounds must capture what probability of containing the true rank of the MRCA? Default 0.95.

Returns

(int, int), optional

Inclusive lower and then exclusive upper bound on estimate or None if no common ancestor between first and second can be resolved with sufficient confidence. (Sufficient confidence depends on confidence_level.)

See Also

calc_rank_of_mrca_uncertainty_between :

Wrapper to report uncertainty of calculated bounds.

calc_rank_of_earliest_detectable_mrca_between :

Could any MRCA be detected between first and second? What is the rank of the earliest MRCA that could be reliably detected?

calc_rank_of_mrca_bounds_provided_confidence_level :

With what actual confidence (i.e., more than requested) is the true rank of the MRCA captured within the calculated bounds?

does_definitively_have_no_common_anestor :

Does the hereditary stratigraphic record definitively prove that first and second could not possibly share a common ancestor?

Notes

The true rank of the MRCA is guaranteed to never fall above the returned bounds but may fall below.

An alternate approach could be to construct the bounds such that the true rank of the MRCA will fall above or below the bounds with equal probability. This would involve setting the confidence level for calculating the first disparity with second to significance_level/2 and the confidence level for calculaing the last comonality with second to 1 - significance_level/2. This means the confidence level applied to calculating the first disparity with second would always be <= 0.5. However, shifting the calculated first disparity with second below the definitive max first retained disparity requires confidence level >= 0.5. So, in practice such a symmetric approach would only result in the lower bound being shifted downward. For this reason, it is no longer provided as an option.

In the absence of evidence to the contrary (i.e., more common strata than spurious differentia collisions alone could plausibly cause), this method assumes no common ancestry between first and second, returning None. This means that if few enough common ranks are shared between first and second (and the differentia bit with is small enough), it may not be possible to detect any common ancestry after accounting for the possibility of spurious differentia collisions (even if common ancestry did exist). So, calls to this method would always return None. Likewise, MRCAs at very early ranks may not be able to be reliably detected due to insufficient evidence. This can lead to cases where columns with true common ancestry have MRCA bounds estimated as None at much higher than the expected failure rate at the given confidence level. Note that with sufficient differentia bit width (i.e., so that even one collision is implausible at the given confidence level) this issue does not occur. Use CalcRankOfEarliestDetectableMrcaWith to determine the earliest rank at which an MRCA could be reliably detected between first and second.