SequenceBouncer
Version: 1.23

----

Author:

Cory Dunn
Institute of Biotechnology
University of Helsinki
Email: cory.dunn@helsinki.fi

----

License:

GPLv3

----

Please cite: 

C.D. Dunn. SequenceBouncer: A method to remove outlier entries from a multiple sequence alignment. bioRxiv. doi:10.1101/2020.11.24.395459.

----

Acknowledgements:

Funding received from the Sigrid Jusélius Foundation contributed to the development of this software. The author thanks Ed Davis, Oregon State University, for suggestions.

----

Requirements:

SequenceBouncer is implemented in Python 3 (current version tested under Python version 3.9.4) 

Dependencies: 

Biopython (tested under version 1.78),
Pandas (tested under version 1.2.4),
Numpy (tested under version 1.20.2),
Matplotlib (tested under version 3.4.2)

----

Usage:

SequenceBouncer.py [-h] -i INPUT_FILE [-o OUTPUT_FILE] [-g GAP_PERCENT_CUT]
                                             [-k IQR_COEFFICIENT] [-n SUBSAMPLE_SIZE] [-t TRIALS]
                                             [-s STRINGENCY] [-r RANDOM_SEED]

Required arguments:
  -i INPUT_FILE, --input_file INPUT_FILE
                        Input multiple sequence alignment file in FASTA format.

Optional arguments:

  -o OUTPUT_FILE, --output_file OUTPUT_FILE
                        Output filename [do not include extensions] (default will be 'input_file.ext').

  -g GAP_PERCENT_CUT, --gap_percent_cut GAP_PERCENT_CUT
                        For columns with a greater fraction of gaps than the selected value, expressed in
                        percent, data will be ignored in calculations (default is 2).

  -k IQR_COEFFICIENT, --IQR_coefficient IQR_COEFFICIENT
                        Coefficient multiplied by the interquartile range that helps to define an outlier
                        sequence (default is 1.0).

  -n SUBSAMPLE_SIZE, --subsample_size SUBSAMPLE_SIZE
                        |> Available for large alignments | The size of a single sample taken from the
                        full dataset (default is entire alignment, but try a subsample size of 50 or 100
                        for large alignments).

  -t TRIALS, --trials TRIALS
                        |> Available for large alignments | Number of times each sequence is sampled and
                        tested (default is to examine all sequences in one single trial, but 5 or 10
                        trials may work well when subsamples are taken from large alignments).

  -s STRINGENCY, --stringency STRINGENCY
                        |> Available for large alignments | 1: Minimal stringency 2: Moderate stringency
                        3: Maximum stringency (default is moderate stringency).

  -r RANDOM_SEED, --random_seed RANDOM_SEED
                        Random seed (integer) to be used during a sampling-based approach (default is
                        that the seed is randomly selected). The user can use this seed to obtain
                        reproducible output and should note it in their publications.

---

This could be the starting point for further exploration of parameters:

i) If the alignment is of moderate size: 

python SequenceBouncer.py -i <input alignment>

ii) If the alignment is of substantial size:

python SequenceBouncer.py -i <input alignment> -n 100 -t 10 -s 2

---

Example files are found at: 

http://doi.org/10.5281/zenodo.4285789