Effects of Substitution Matrices, Trimming and Outliers on Alignments and Phylogenetic Tree Constructions of Big Dataset

The substitution matrix is important for the multiple sequence alignment (MSA). Thereafter the MSA is often subject to trimming and eliminating of detected outlier sequences. However, the effects of substitution matrices, methods for trimming and eliminating outlier sequences in larger datasets require additional attention. In this study, (1) a big dataset of 3228 sequences were aligned using the R language package msa with BLOSUM30, BLOSUM40, BLOSUM50, BLOSUM65, BLOSUM80 and Gonnet substitution matrices; (2) the alignments were trimmed using the software package trimA1 with three automation methods, gappyout, strict and strictplus, and were detected for outliers using the R language package OD-seq with affine and linear methods; and (3) 120 phylogenetic trees were built from untrimmed and trimmed alignments. The phylogenetic information, sequence logo, alignment statistics, the normalized Robinson-Foulds distance, locality of a reference sequence, and identical sequences were used for analyses. The results demonstrate the complicated effects on alignments and phylogenetic tree constructions. The relevant implications were discussed in detail.