Extended Data Fig. 3: Robustness of evolutionary lineage reconstruction with sequence data noise.
From: Deep distributed computing to reconstruct extremely large lineage trees
The scatter plots represent accuracies to reconstruct the SILVA 16S rRNA lineage from the input sequence datasets including different fractions of noise sequences. The noise sequences were generated by shuffling nucleotide positions of randomly selected sequences by keeping their alignment gap positions in the multiple sequence alignment result. The lineage dendrograms represent the ones reconstructed from the sequence dataset with 20% noise using FRACTALized RapidNJ, RAxML and FastTree.
