Different amount of PD between 2 runs

[louisehendrickx]

These 2 phylogenies are with and without Wollemia nobilis. We ran phylonext for both of these phylogenies. All parameters were default except for type of phylogenetic tree labels (Latin) family (Araucariaceae), min year (1980) and country (Australia), number of randomisation (20).

The problem is that some cells where Wollemia nobilis is not present have different amounts of PD between the 2 runs. We would expect if Wollemia nobilis is not present, they would have exactly the same amount of PD.

These are the two phylogenies we ran:

(((Agathis_australis,Agathis_palmerstonii,((((Agathis_atropurpurea,Agathis_microstachya),Agathis_robusta),((Agathis_dammara,Agathis_borneensis),(Agathis_silbae,Agathis_macrophylla))),((Agathis_moorei,(Agathis_ovata,Agathis_corbassonii)),Agathis_lanceolata)),Agathis_montana,Agathis_obtusa,Agathis_vitiensis,Agathis_brownii,Agathis_alba,Agathis_philippinensis,Agathis_flavescens,Agathis_orbicula,Agathis_kinabaluensis,Agathis_labillardierei,Agathis_lenticula,Agathis_endertii,Agathis_silbaii)Agathis,(Wollemia_nobilis)Wollemia),(((Araucaria_heterophylla,((((Araucaria_columnaris,(Araucaria_luxurians,Araucaria_nemorosa)),((Araucaria_humboldtensis,Araucaria_laubenfelsii),Araucaria_rulei)),((Araucaria_bernieri,(Araucaria_schmidii,Araucaria_subulata)),Araucaria_scopulorum)),(Araucaria_biramulata,Araucaria_montana))),(Araucaria_cunninghamii_var._cunninghamii)Araucaria_cunninghamii),((Araucaria_araucana,Araucaria_angustifolia),(Araucaria_bidwillii,Araucaria_hunsteinii)),Araucaria_excelsa,Araucaria_muelleri,Araucaria_braziliana,Araucaria_lignitici,Araucaria_bernieri,Araucaria_annulata,Araucaria_balcombensis,Araucaria_fimbriatus,Araucaria_grandifolia,Araucaria_hastiensis,Araucaria_longifolia,Araucaria_marensii,Araucaria_nathorsti,Araucaria_planus,Araucaria_prominens,Araucaria_readiae,Araucaria_taeriensis,Araucaria_uncinatus,Araucaria_araucoensis,Araucaria_brownii,Araucaria_jeffreyi,Araucaria_marylandica,Araucaria_neocookii,Araucaria_pichileufensis,Araucaria_prodromus)Araucaria,(Agathoxylon_pseudoparenchymatosum)Agathoxylon,(Balmeisporites_minutus)Balmeisporites)Araucariaceae;

((Agathis_australis,Agathis_palmerstonii,((((Agathis_atropurpurea,Agathis_microstachya),Agathis_robusta),((Agathis_dammara,Agathis_borneensis),(Agathis_silbae,Agathis_macrophylla))),((Agathis_moorei,(Agathis_ovata,Agathis_corbassonii)),Agathis_lanceolata)),Agathis_montana,Agathis_obtusa,Agathis_vitiensis,Agathis_brownii,Agathis_alba,Agathis_philippinensis,Agathis_flavescens,Agathis_orbicula,Agathis_kinabaluensis,Agathis_labillardierei,Agathis_lenticula,Agathis_endertii,Agathis_silbaii)Agathis,(((Araucaria_heterophylla,((((Araucaria_columnaris,(Araucaria_luxurians,Araucaria_nemorosa)),((Araucaria_humboldtensis,Araucaria_laubenfelsii),Araucaria_rulei)),((Araucaria_bernieri,(Araucaria_schmidii,Araucaria_subulata)),Araucaria_scopulorum)),(Araucaria_biramulata,Araucaria_montana))),(Araucaria_cunninghamii_var._cunninghamii)Araucaria_cunninghamii),((Araucaria_araucana,Araucaria_angustifolia),(Araucaria_bidwillii,Araucaria_hunsteinii)),Araucaria_excelsa,Araucaria_muelleri,Araucaria_braziliana,Araucaria_lignitici,Araucaria_bernieri,Araucaria_annulata,Araucaria_balcombensis,Araucaria_fimbriatus,Araucaria_grandifolia,Araucaria_hastiensis,Araucaria_longifolia,Araucaria_marensii,Araucaria_nathorsti,Araucaria_planus,Araucaria_prominens,Araucaria_readiae,Araucaria_taeriensis,Araucaria_uncinatus,Araucaria_araucoensis,Araucaria_brownii,Araucaria_jeffreyi,Araucaria_marylandica,Araucaria_neocookii,Araucaria_pichileufensis,Araucaria_prodromus)Araucaria,(Agathoxylon_pseudoparenchymatosum)Agathoxylon,(Balmeisporites_minutus)Balmeisporites)Araucariaceae;

[shawnlaffan]

If the reported PD index is PD_P then that is the proportion of the tree in each sample, calculated as the PD score divided by the sum of all branch lengths. This could explain the differences.

See https://github.com/shawnlaffan/biodiverse/wiki/Indices#phylogenetic-diversity for index names and formulae.

[louisehendrickx]

I think I am not looking at PD_P because the index is showing the absolute values of PD, not the percentages. I attached 2 screenshots of the result we get when we run the mentioned phylogenies. The PD shows different values, although Wollemia nobilis does not occur there (see third screenshot). This is the case for many of the hexagons.

[shawnlaffan]

I think I am not looking at PD_P because the index is showing the absolute values of PD, not the percentages.

The PD legend is in percentages, though, which suggests PD_P. @vmikk can you clarify?

In terms of the spatial distribution it would appear that the Wollemia data are not being cleaned correctly. Only one of the hexagons in the third plot should contain it. Several appear to be associated with capital cities and thus probably are herbaria and/or botanic gardens.