/FastLoo

Subquadratic leave-one-out outlier detection via a APPLES-2

Primary LanguagePython

FastLoo

Introduction

This is a subquadratic algorithm for detecting backbone sequences that are discordant with a given input phylogeny. A usecase is when the input tree is computed using a seperate set of biological sequences on the same set of species. More concretely, the input tree can be species tree inferred using many genes and input alignment can be a single gene.

The algorithm is subquatratic with respect to the number of species thanks to APPLES-2.

Usage

The first input is a multiple sequence alignment. The second input is a phylogenetic tree whose leaves are the input alignments. The third argument is nuc for nucleotide and prot for protein.

Examples:

For nucleotide alignment:

FastLoo aln.fa backbone.nwk nuc

For protein alignment:

FastLoo aln.fa backbone.nwk prot

You can test the tool by running the following command

FastLoo data/ref.fa data/backbone.nwk nuc

Installation:

run install.sh and then activate the conda environment conda activate fastloo