Figure 3From: Parametric and non-parametric masking of randomness in sequence alignments can be improved and leads to better resolved treesNeighbor-Net graphs. Neighbor-Net graphs generated with SplitsTree 4.10 based on concatenated supermatrices of unmasked (left), ALISCORE masked (middle), and GBLOCKS(none) masked (right) data. mtI, mtII and EST networks depend on a T-COFFEE alignment, the 12S + 16S rRNA network on a PCMA alignment. Neighbor-Nets were calculated with uncorrected p-distances. All inferred Neighbor-Net graphs are given in the additional File 1. Tree like structures in these graphs indicate distinct signal-like patterns in the corresponding alignment. Graphs generated from ALISCORE data sets are more tree-like. Lack of information leads to star-like graphs, conflicting signal produces cobwebs.Back to article page