Sequence Extraction

SEGUL can extract sequences based on the sequence ID in a collection of alignments. You can input the sequence ID in three ways:

1. Using the standard input (stdin) via terminal
2. By inputting a list of IDs in a text file
3. Using regular expression

By default, segul will write the results in a nexus format and use the name from the original input. More details below.

Extracting sequences using stdin

This method is helpful if you only extract a few sequences across all your alignments. To input the ID list, use --id option. Semi-colons should separate the list of IDs.

segul sequence extract <input-option> [alignment-dir] --id="[list-of-sequence-id]"

For example, using the command below, we will extract seq_1, seq_2, and seq_3 from our alignments stored in alignments/ directory:

segul sequence extract --dir alignments/ --id="seq_1;seq_2;seq_3"

Caution

To input a list of IDs, you must use the equal sign = and quotation marks ".

Extracting sequences using an input file

With a long list of IDs, it will be more convenient to input the list of the IDs instead. segul --file option is available for this task. You only need to list all the IDs in a text file with each ID in a new line as below:

seq_1
seq_2
seq_3
seq_4
seq_5

The command will be:

segul sequence extract --dir alignments/ --file id.txt

For lots of ID, you can generate all the IDs first using SEGUL sequence ID extraction. Open the text file using your text editor and remove the sequence IDs you don't need. Use this file as an input for segul sequence extract.

Extracting sequences using regular expression (REGEX)

This is the most flexible option to extract sequences from a collection of alignments. SEGUL uses the Rust regex library to parse regular expression syntax. The syntax is similar to Perl-style regex syntax (details here). To test your syntax, you can use a website like https://regex101.com/.

Use the option --re= for a regular expression and put the syntax in a single or double quotation. For example, our sequence ID is named this way genus_species_voucherNo. We want to extract the sequences from the same genus: Mus. We could write a regular expression that searches for sequence ID that starts with Mus as below. The ^ indicates the start of the word in REGEX syntax.

segul sequence extract --dir alignments/ --re="^Mus"

We could also search it with an insensitive match by adding (?i) syntax.

segul sequence extract --dir alignments/ --re="^(?i)mus"

Each alignment containing matched sequences will be written to the output files. The alignment's name will remain the same as the original.

Specifying output directory and format

By default, the output files will be written into Sequence-Extract directory. To change the directory name, use the --output or -o option. For the output files, the matched sequences will be written in nexus format using the alignment original name. You can change the output format using the --output-format (or -F in the short version) option. For example, below, we'll be able to change both the output directory and the output format of the resulting files.

segul sequence extract --dir alignments/ --re="^Mus" --output Mus-alignment/ --output-format fasta

All options

Option	Description
-d, --dir	Directory with alignments
-i, --input	Alignment files. Allows wildcards
-f, --input-format	Input format: auto (default), fasta, nexus, or phylip
--datatype	Data type: dna (default), aa, or ignore
--id=	Sequence IDs to extract, separated by semicolons
--file	Text file of sequence IDs to extract, one per line
--re=	Regular expression that matches the sequence IDs to extract
-F, --output-format	Output format: nexus (default), phylip, fasta, nexus-int, phylip-int, or fasta-int
-o, --output	Output directory. Default: Sequence-Extract
--force	Overwrite an existing output directory
--log	Log file path. Default: segul.log