Limits...
Concatabominations: identifying unstable taxa in morphological phylogenetics using a heuristic extension to safe taxonomic reduction.

Siu-Ting K, Pisani D, Creevey CJ, Wilkinson M - Syst. Biol. (2014)

View Article: PubMed Central - PubMed

Affiliation: Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK agalychnica@gmail.com.

AUTOMATICALLY GENERATED EXCERPT
Please rate it.

Perhaps less controversial is that individual taxa may sometimes be relatively phylogenetically unstable by virtue of limited data and extensive missing data (e.g., ; ; ; ). developed an approach for diagnosing taxon instability due to missing data a priori termed safe taxonomic reduction (STR)... STR allows the identification of “rogue” taxa that can be removed from a data set safe in the knowledge that their removal will not impact upon the interrelationships that will be inferred among the remaining taxa under the parsimony criterion... The potential benefits of such deletion are reductions in numbers of optimal trees and run times and better resolved consensus summaries... In such cases there may be many pairs of leaves with p-distances of zero but, because of the distribution of missing entries, the character states of neither are a proper subset of those of the other (category D, Fig. 1). called such pairs of leaves “potential taxonomic equivalents that are asymmetric both ways” (we will call them pairs) and recognized that, in contrast to the other categories of taxonomic equivalence, the deletion of either member of the pair cannot be guaranteed to be safe a priori... We also define, for each leaf, ABC as the number of taxonomic equivalents of that leaf in the STR categories, , or (each of which identifies scope for a priori safe deletion)... Note however that this improvement of the strict consensus can be obtained through the deletion of just Hulsanpes and Saurornitholestes... Although deletions of Liliensternus and/or, Procompsognathus are both safe and reduce the number of MPTs they are not effective at increasing the resolution of the corresponding strict consensus... The two leaves with the highest * (Hulsanpes and Sauronitholestes) scores are also identified by traditional STR as taxa that can be safely deleted... Deletion of Hulsanpes alone reduces the number of MPTs for the remaining data to 45,654 without affecting tree length but does not improve (increase the number of splits in) the corresponding strict consensus... Whereas STR identifies two additional taxa (Procompsognathus and Liliensternus) that can also be safely deleted, ranking based on * scores prompts the experimental deletion of Coelurus... As already noted, the deletion of Procompsognathus and Liliensternus reduces the number of MPTs (to 197) but does not further improve the strict consensus... One undoubted attraction of STR is that a taxon is deleted a priori only if we are certain that this deletion cannot impact upon the relationships inferred among the remaining taxa... In any particular case there may be useful safe taxon deletions that are not identified a priori using STR... As highly connected, potentially unstable, taxa are deleted any changes in the degree of the remaining vertices and of their relative rankings will be apparent... Natural stopping points for experimental deletion are when formerly connected clusters of taxa completely separate or when connected taxa cannot be safely deleted or their safe deletion does not improve the consensus.

Show MeSH
Producing a concatabomination () for a  pair of taxa with asymmetric potential equivalence both ways. Arrows show how the concatabomination leads to a composite taxon with missing data of each original taxon replaced where possible by data from its pair. In other words, the concatabomination of a  pair is a taxon comprising the union of the character states of the  pair.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4265141&req=5

Figure 2: Producing a concatabomination () for a pair of taxa with asymmetric potential equivalence both ways. Arrows show how the concatabomination leads to a composite taxon with missing data of each original taxon replaced where possible by data from its pair. In other words, the concatabomination of a pair is a taxon comprising the union of the character states of the pair.

Mentions: Our approach to determining whether homoplasy is increased by forcing leaves to go together makes use of compatibility methods (e.g., Meacham and Estabrook 1985). Two characters are compatible if there is some tree on which they can both fit without any extra steps (homoplasy) and simulations have shown that compatibility decreases as homoplasy increases both for whole matrices (O'Keefe and Wagner 2001) and individual characters (Wagner 2012). We count the total number of character pairs in the data that are incompatible (Le Quesne 1969) and use this as a proxy estimate of homoplasy in the original data. We then combine the data for a pair of leaves to make what we call a “concatabomination” (Fig. 2), add this construct to the original data, and recalculate the pairwise incompatibility. We repeat the latter for each pair in turn. For each leaf, we define * as the number of times that leaf contributes to a concatabomination that does not appear to increase homoplasy (i.e., does not increase the number of pairwise character incompatibilities) in the data. We also define, for each leaf, ABC as the number of taxonomic equivalents of that leaf in the STR categories , , or (each of which identifies scope for a priori safe deletion). Taxa can be ranked based on these individual scores or their sum.


Concatabominations: identifying unstable taxa in morphological phylogenetics using a heuristic extension to safe taxonomic reduction.

Siu-Ting K, Pisani D, Creevey CJ, Wilkinson M - Syst. Biol. (2014)

Producing a concatabomination () for a  pair of taxa with asymmetric potential equivalence both ways. Arrows show how the concatabomination leads to a composite taxon with missing data of each original taxon replaced where possible by data from its pair. In other words, the concatabomination of a  pair is a taxon comprising the union of the character states of the  pair.
© Copyright Policy - creative-commons
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4265141&req=5

Figure 2: Producing a concatabomination () for a pair of taxa with asymmetric potential equivalence both ways. Arrows show how the concatabomination leads to a composite taxon with missing data of each original taxon replaced where possible by data from its pair. In other words, the concatabomination of a pair is a taxon comprising the union of the character states of the pair.
Mentions: Our approach to determining whether homoplasy is increased by forcing leaves to go together makes use of compatibility methods (e.g., Meacham and Estabrook 1985). Two characters are compatible if there is some tree on which they can both fit without any extra steps (homoplasy) and simulations have shown that compatibility decreases as homoplasy increases both for whole matrices (O'Keefe and Wagner 2001) and individual characters (Wagner 2012). We count the total number of character pairs in the data that are incompatible (Le Quesne 1969) and use this as a proxy estimate of homoplasy in the original data. We then combine the data for a pair of leaves to make what we call a “concatabomination” (Fig. 2), add this construct to the original data, and recalculate the pairwise incompatibility. We repeat the latter for each pair in turn. For each leaf, we define * as the number of times that leaf contributes to a concatabomination that does not appear to increase homoplasy (i.e., does not increase the number of pairwise character incompatibilities) in the data. We also define, for each leaf, ABC as the number of taxonomic equivalents of that leaf in the STR categories , , or (each of which identifies scope for a priori safe deletion). Taxa can be ranked based on these individual scores or their sum.

View Article: PubMed Central - PubMed

Affiliation: Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK Department of Biology, National University of Ireland, Maynooth, Co. Kildare, Ireland; School of Biological Sciences, University of Bristol, Bristol BS8 1TQ, UK; Department of Life Sciences, The Natural History Museum, London SW7 5BD, UK; and Institute of Biological, Environmental & Rural Sciences, Aberystwyth University, Aberystwyth SY23 3FG, UK agalychnica@gmail.com.

AUTOMATICALLY GENERATED EXCERPT
Please rate it.

Perhaps less controversial is that individual taxa may sometimes be relatively phylogenetically unstable by virtue of limited data and extensive missing data (e.g., ; ; ; ). developed an approach for diagnosing taxon instability due to missing data a priori termed safe taxonomic reduction (STR)... STR allows the identification of “rogue” taxa that can be removed from a data set safe in the knowledge that their removal will not impact upon the interrelationships that will be inferred among the remaining taxa under the parsimony criterion... The potential benefits of such deletion are reductions in numbers of optimal trees and run times and better resolved consensus summaries... In such cases there may be many pairs of leaves with p-distances of zero but, because of the distribution of missing entries, the character states of neither are a proper subset of those of the other (category D, Fig. 1). called such pairs of leaves “potential taxonomic equivalents that are asymmetric both ways” (we will call them pairs) and recognized that, in contrast to the other categories of taxonomic equivalence, the deletion of either member of the pair cannot be guaranteed to be safe a priori... We also define, for each leaf, ABC as the number of taxonomic equivalents of that leaf in the STR categories, , or (each of which identifies scope for a priori safe deletion)... Note however that this improvement of the strict consensus can be obtained through the deletion of just Hulsanpes and Saurornitholestes... Although deletions of Liliensternus and/or, Procompsognathus are both safe and reduce the number of MPTs they are not effective at increasing the resolution of the corresponding strict consensus... The two leaves with the highest * (Hulsanpes and Sauronitholestes) scores are also identified by traditional STR as taxa that can be safely deleted... Deletion of Hulsanpes alone reduces the number of MPTs for the remaining data to 45,654 without affecting tree length but does not improve (increase the number of splits in) the corresponding strict consensus... Whereas STR identifies two additional taxa (Procompsognathus and Liliensternus) that can also be safely deleted, ranking based on * scores prompts the experimental deletion of Coelurus... As already noted, the deletion of Procompsognathus and Liliensternus reduces the number of MPTs (to 197) but does not further improve the strict consensus... One undoubted attraction of STR is that a taxon is deleted a priori only if we are certain that this deletion cannot impact upon the relationships inferred among the remaining taxa... In any particular case there may be useful safe taxon deletions that are not identified a priori using STR... As highly connected, potentially unstable, taxa are deleted any changes in the degree of the remaining vertices and of their relative rankings will be apparent... Natural stopping points for experimental deletion are when formerly connected clusters of taxa completely separate or when connected taxa cannot be safely deleted or their safe deletion does not improve the consensus.

Show MeSH