Commandline Reference

SINA adds sequences to an existing multiple sequence alignment (MSA). It can also execute a homology search based on the computed alignment and generate a per sequence classifications from the search results.


sina [options] -i <unaligned> -r <reference> -o <aligned>

sina [ -h | --help | --help-all | --version | --has-cli-vers ]


sina aligns the sequences in the file <unaligned> to match the alignment in <reference> and places the aligned sequences in the file <aligned>.

Please refer to the publication for details on the algorithm(s).

General Options

-h, --help

Displays brief command line description.

-H, --help-all

Displays full command line description. If in doubt, refer to this as it will always be in sync with your installation of SINA.

-V, --version

Shows the version of SINA.

-i filename, --in=filename (-)

Specifies the file containing the input sequences. Allowable file formats are ARB (see ARB Options) and FASTA (see FASTA Options, optionally gzipped). The format will be selected based on the file name unless overridden with :option:–intype`.

Special file names: “-” (dash) will read sequences from standard input. “:” (colon) will connect to a running ARB database (SINA must be a child process of ARB).

-o filename [filename [...]], --out=filename (-)

Specifies the output file(s) to which the aligned sequences and/or meta data will be written. Allowable file formats are ARB (see ARB Options) and FASTA (see FASTA Options, optionally gzipped) and CSV (see CSV Options). The format will be selected based on the file name unless overridden with --outtype. Specifying multiple names or specifying the option multiple times will output all data to each file.

See -i for special filenames - and :. Use -o /dev/null to disable output.

Not specifying this option at all will write sequences to stdout if the input is FASTA format. When reading from an ARB database, output is written back to the source database (specified with -i).

-r filename, --db=filename

Specifies the file containing the reference alignment. This file must be in ARB format.

To convert a reference alignment from FASTA to ARB format, run:

sina -i reference.fasta --prealigned -o reference.arb
-t [all], --turn [=all]

Enables turn check stage. Sequences not oriented in accordance with the reference database will be reverse complemented as needed.

If all is specified, sequences will also be tested for only reversal or only complemented (this should only be necessary if your data was mishandled).

-S, --search

Enables the search stage. See Search & Classify Options below for more information.

-P, --prealigned

Disables the alignment stage. This is useful if you have already aligned sequences you wish to pass directly into the search stage, or if you want to use SINA to convert between any of its supported file formats.

-v, --verbose

Increase logging verbosity. Can be specified multiple times.

-q, --quiet

Decrease logging verbosity. Can be specified multiple times.


Specify log file. The output written to the log file will always be verbose and is not affected by using -v or -q.


Configures how meta data (such as alignment score or sequence classification results) are to be exported.

No output other than in the log is generated.
Appends meta data as [key=value] pairs to the FASTA header line
Appends meta data as ; key: value lines between the FASTA header and the sequence data.
Writes meta data into a CSV side car file.

Deprecated since version 1.7.0: Use -o output.csv instead.

-f fields, --fields=fields

Configures the set of fields written to the output file. See Field Reference for a description of the available fields.

-p, --threads (automatic)

Override automatic detection of the number of threads used by SINA. This is usually only necessary if you need to constrain SINA to a lower number of threads. According to the Intel engineers whose Threaded Building Blocks library does the thread number detection for SINA, the only reason to use this parameter should be scalability testing.

--num-pts (1)

Set the maximum number of ARB PT server instances used by SINA. See also --fs-engine below. If you are using the pt-server engine, this setting will be the limiting factor in your throughput. Be aware, however, that each PT server will occupy additional system memory. Choosing a too high value may cause SINA to fail with out-of-memory errors.

--add-relatives=n (0)

Add up to n reference sequences for each query sequence to the output file. If Search & Classify is enabled via --search, the reference sequences are selected from the search result. Otherwise, they are selected from the query’s alignment reference set.

If the source set is smaller than n, no further sequences are added to the output. Sequences already included are skipped, but count towards the n of the query sequence.

Reference Selection Options

These options configure how the set of reference sequences used during alignment is selected from the configured reference database.


Selects the search engine used to find closely related reference sequences for the alignment stage.

Uses the ARB PT server to execute the k-mer search. The ARB PT server is a truncated suffix trie implementation implemented as part of the ARB package.
Uses an internal k-mer search implementation.
--fs-kmer-len=k (10)

Set the size of k for the reference search. For SSU rRNA sequences, the default of 10 is a good value. For different sequence types, different values may perform better. For 5S, for example, 6 has shown to be more effective.

--fs-min=n (15)

Set the minimum number of reference sequences used for each query.

--fs-max=n (40)

Set the maximum number of reference sequences used for each query.

--fs-msc=n (0.7)

Set the minimum similarity reference sequences are required to have with the query sequence. This affects the range between --fs-min and --fs-max.

--fs-req=n (1)

Set the minimum number of reference sequences that must be found in order to attempt alignment. If fewer sequences than indicated here are found, the respective query sequence will be discarded.

--fs-req-full=n (1)

Set the minimum number of full length (see --fs-full-len) reference sequences that must be included in the selected reference set. The search will proceed regardless of other settings until this setting has been satisfied. If it cannot be satisfied by any sequence in the reference database, the query sequence will be discarded.

This setting exists to ensure that the entire length of the query sequence will be covered in the presence of partial sequences contained within your reference database.

If you are working with sequences other than 16S, you need to adjust this value or the value of --fs-full-len accordingly. In particular when working with short reference sequences, this setting may prevent any acceptable reference sequences from being found, leading to no sequences being aligned.
--fs-full-len=n (1400)

Set the minimum length a sequence is required to have to be considered full length.

--fs-req-gaps=n (10)

Set the minimum number of gaps a reference sequence is required to contain to be considered. This setting ensures that unaligned sequences contained within the reference database are not used as reference (this may happen when SINA is used from within ARB).

--fs-min-len=n (150)

Set the minimum length reference sequences are required to have. Sequences shorter than this will not be included in the selection.

If you are working with particularly short reference sequences, you will need to lower this settings to allow any reference sequences to be found.

Search & Classify Options

When enabled via --search, SINA will execute a homology search. Unlike most homology search tools, SINA uses the inferred multiple sequence alignment to determine the similarity of each query with the reference sequences, rather than computing pairwise optimal alignments. The similarity values will therefore be generally lower than the results of a pairwise alignment based homology search.

Based on the search results, SINA can be instructed to compute a lowest common ancestor (LCA) based classification of the input sequences. For this, your reference database must include a field containing taxonomic classifications for each reference sequence. The field contents must be in the format t Domain;Phylum;…. SINA will compute query classifications as the deepest classification shared by at least the fraction --lca-quorum of the search result.

--search-db=filename (=db)

Specify an alternate reference database to use for search and classify. This can be useful if you have a specially curated alignment reference, but wish to search a larger set of sequences for classification purposes.


Override the value of --fs-engine for use within the search module.

--search-min-sim=id (0.7)

The minimum fractional identity each result sequence must have with the query.

--search-max-result=n (10)

The maximum number of search results to return for each query sequence.


Enables the classification stage. The parameter name must be a colon or comma separated list of field names in the search database containing the classification reference data. Use --arb-list-fields to show a list of the fields available in a given ARB database. When using a SILVA ARB database as reference, the fields tax_slv, tax_embl (renamed to tax_embl_ebi_ena in newer releases) and tax_ltp contain the reference classifications according to the SILVA, EMBL-EBI/ENA and LTP taxonomies, respectively. When using a SILVA SSU ARB database, the fields tax_gg (only older databases) and tax_rdp are available additionally, containing the reference classifications according to RDP II and Greengenes, respectively. Newer SILVA databases also contain the field tax_gtdb containing classifications from the Genome Taxonomy Database.

--lca-quorum=fraction (0.7)

Sets the fraction of the search result that must share the same classification. Using the default parameters --search-max-result=10 and --lca-quorum=0.7, this means that the deepest classification shared by 7 out of the top 10 search results is chosen for the query sequence.

Advanced Options


Print the values of all configuration options (including defaults) at startup.

--intype=[auto|arb|fasta] (auto)

Set the file format for --in. If set to auto (default), the type is selected based on the file extension.

--outtype=[auto|arb|fasta|csv|none] (auto)

Set the file format for --out. If set to auto (default), the type is selected based on the file extension. The option can be specified multiple times. It applies to all files listed in the next --out option. If no output files are specified and this option is set to none, no output is produced at all.


Preserve the order of the input sequences in the output.

--max-in-flight=n (2 * number of CPU cores)

Set the maximum number of sequences “in flight”, i.e. processed in parallel.


Verify that this version of SINA supports the CLI version cliversion. Exits immediately with exit code 0 if true and 1 if false.


Backwards compatibility alias for --prealigned.

Logging Options


Show differences between the inferred alignment and the original alignment. Requires either aligned sequences to be passed into sina via --in or that a database with matching names is specified using --orig-db.


Fix text below

This flag enables visualization of alignment differences. This feature allows you to quickly assess where your alignment differs from the one SINA computed. By also showing you the alignment of the reference sequences used for aligning the sequence, you can get an idea of why SINA came to its conclusions. Many cases of “sub-optimal” alignment can be attributed to inconsistent alignment of the reference sequences. To fix such problems, you could either correct the alignment of the reference sequences or add your corrected sequence to the reference alignment.

Alignment difference visualization requires that the input sequences be already aligned in a way compatible with the used reference alignment. For positions at which the original alignment and the alignment computed by SINA differ, output as shown below will be printed to the log:

Dumping pos 1121 through 1141:
---------  4 14 16-17 21 24
G-C-AGUC-  40 <---(%% ORIG %%)
GCA--GUC-  41 <---(## NEW ##)
GCA-AGUC-  0-3 5-13 15 18-20 22-23 25-27 29-39

In this case, the bases ‘texttt{C}’ and ‘texttt{A}’ where placed in other columns than as per the original alignment. The original alignment is marked with texttt{<-{}–(%% ORIG %%)}. The new alignment is marked with texttt{<-{}–(## NEW ##)}. The numbers to the right of the alignment excerpt indicate the indices of the sequences in the alignment reference (field


Show distance to original alignment


describe values generated


Specify a database containing the original alignments for use with --show-dist and --show-diff. The sequence names in the input file and in the reference database must match exactly.


Use ANSI codes to show alignments dumped by --show-diff in color.

ARB Options

These options configure behavior supported only by the ARB backend for input and output sequences.


Show the per-sequence meta-data fields available in the specified ARB database and exit.


Set Mark on sequences copied from the reference.


Set Mark on sequences updated or added by alignment stage.

--prot-level=n (4)

Set the protection level to use when writing sequences to the output database.


Instead of iterating over the entire input database, process only the sequences listed in filename. The names must match the ARB name field and be separated by newlines. Use “-” to read from standard input.

--select-step=n (1)

Process only every nth sequence. Can be combined with --select-file and --select-skip.

--select-skip=n (0)

Do not process the first n sequences. Can be combined with --select-file and --select-step.

FASTA Options

These options configure behavior supported only by the FASTA backend for input and output sequences.

--line-length=n (0)

Output sequences using at most n characters per line. Set to 0 to place the entire output sequence on one line.


Exclude sequences sharing less than id fractional identity with any of the alignment reference sequences from the output. Implies --calc-idty.


Write output sequences as DNA, rather than the default RNA. (I.e. use T and t rather than U and U).


Use dots (“.”) rather than dashes (“-“) for gaps that indicate missing data rather than an actual insertion/deletion. Most often, those are only the terminal gaps at the ends of the alignment.


Check whether internal dots are handled correctly.


Only process sequences starting withing the nth block of bytes within the input FASTA file. The first block has index 0.


Sets the size in bytes for the blocks used by --fasta-idx.

CSV Options

These options configure the CSV output backend.


Enables RFC4180 compliant CSV output using CRLF as line separator. This was the default behavior for the old, deprecated CSV writer enabled using --meta-fmt.

--csv-sep=sep (,)
Specifies the string used to separate fields. May be multple characters. By default, a comma is used when writing to STDOUT or files ending in .csv or .csv.gz and a TAB character is used when writing to files ending in .tsv or .tsv.gz.
--csv-id=id (name)

Specifies the field name used for the always present ID column. For sequences read from FASTA, this is the first word of the header. For sequences read from ARB, this is the name field. By default, the column name is name.

Alignment Options


Forces computing the alignment of query sequences even if a reference sequence containing the exact sequence was found. Without this flag, SINA will copy the alignment from the reference sequence.

--overhang=[attach|remove|edge] (attach)

Configures how unaligned bases at the edge of the alignment (overhanging bases) should be handled.

Overhang bases will be placed next to the last aligned base consecutively.

Overhang bases will be deleted.


This feature appears to be broken.

Overhang bases will be placed next to the outer edge of the alignment.

Configures which bases should be written using lower case characters.

All bases will use upper case characters
All bases will be written using the case they had in the input data.
Aligned bases will be written in upper case; unaligned bases will be written in lower case. This serves to mark sections of the query sequences that could not be aligned because they were insertions (internal or edge) with respect to any of the reference sequences.

Configures how the alignment width is preserved.


The alignment is executed without constraining insertion sizes. Insertions for which insufficient columns exist between the adjoining aligned bases are force fitted into the alignment using NAST. That is, the minimum number of aligned bases to the left and right of the insertion are moved to accommodate the insertion.

This mode will add warnings to the log for each sequence in which aligned bases had to be moved.


The alignment is executed using a scoring scheme disallowing insertions for which insufficient columns exist in the alignment.

This mode causes less “misalignments” than the shift mode as it computes the best alignment under the constraint that no columns may be added to the alignment. However, it will not show if the computed alignment suffered from a lack of empty columns.


The alignment is executed without constraining insertion sizes. Insertions larger than the number of columns between the adjoining aligned bases are truncated.

While this mode yields the most accurate alignment for sequences with large insertions, it should be used with care as it modifies the original sequence.


Instructs SINA to use a profile vector instead of a DAG to perform the alignment. That is, the base frequencies for all selected reference sequences are collected into a vector and the query is aligned to this vector weighting the alignment scores according to the respective frequencies.

This feature was added in response to the requests of a reviewer of the original SINA publication and only intended to demonstrate that the DAG/POA approach is superior to the profile vector approach. Do not use this other than for testing.

--fs-weight=weight (1)

Adjust the weight factor for the frequency at which a node was observed in the reference alignment. Use 0 to disable weighting.

This feature prefers the more common placement for bases with inconsistent alignment in the reference database.

--match-score=n (2)

Configures the score given for a match (should be positive).

--mismatch-score=n (-1)

Configures the score given for a mismatch (should be negative).

--pen-gap=n gap open penalty (5)

Configures the penalty subtracted from the score for opening a gap (should be positive).


Configures the penalty subtracted from the score for extending a gap (should be positive).


Writes the DAG computed from the reference sequences for each query sequences to disk in dot format.


Weights the match and mismatch scores according to the overall base frequencies observed in the database.

This feature is experimental and does not currently improve the results.


Writes the names of the alignment reference sequences into the field used_rels. This option allows using the ARB mark used rels feature to highlight the reference sequences used to align a given query sequence.


Computes the highest similarity the aligned query sequence has with any of the sequences in the alignment reference set. The value is written to the field align_ident_slv.

Advanced Reference Selection Options


Alias of --db for backwards compatibility.

--ptport=port_or_socket (:/tmp/sina_pt_<pid>)

Configures the port or socket on which the ARB PT server for the reference alignment is expected or started. To use a TCP port, specify <hostname>:<port>. If <hostname> is not localhost, the PT server must be launched externally. To use a Unix socket, specify :<filename>.

When --num-pts is greater than 1, the additional PT servers port names are generated by appending the respective number. Using port numbers greater of equal to 10000 will therefore not work.

By default, the file /tmp/sina_pt_<pid> is used, where <pid> is replaced by the process ID of the SINA instance.


Use all k-mers occurring in the query sequence in the search. By default, only k-mers starting with an A are used for extra performance.

--fs-kmer-mm=n (0)

Allow k-mer matches in the reference database to contain n mismatches. This feature is only supported by the pt-server search engine and requires substantial additional compute time (in particular for n > 1).


Use absolute (number of shared k-mers) match scores in the kmer search rather than relative (number or shared k-mers divided by length of reference sequence) match scores.

--fs-msc-max=id (2)

Overriding all other options, reference sequences having a similarity with the query higher than this value are excluded from the alignment reference.

This option artificially increases the difficulty of the alignment by increasing the distance of a query to any reference found in the database. It’s purpose of this option is to generate a sufficiently large N of test cases for statistical analysis of SINA’s accuracy for sequences distant to the reference alignment.


Excludes sequences from the alignment reference sharing the same name as the respective query sequence. (For testing and evaluation).

--gene-start=n (0)

Sets the beginning of the gene within the reference alignment. See --fs-cover-gene.

--gene-end=n (0)

Sets the end of the gene within the reference alignment. See --fs-cover-gene.

--fs-cover-gene=n (0)

Similar to --fs-req-full, this option requires a total of n sequences to cover each the beginning and the end of the gene within the alignment. This option is more precise than --fs-req-full, but requires that the column numbers for the range in which the full gene is expected be specified via --gene-start and --gene-end.


Chooses an ARB posvar filter to use for weighting alignment positions by their variability.


Configures a database field using which the value of --filter is determined by majority vote from the selected reference sequences. Since the filters are usually computed at domain level, this approach is usually sufficient to select an appropriate filter. For SILVA database, the field tax_slv contains appropriate data.

--auto-filter-threshold arg

Sets the minimum quorum required for automatic filter selection. See --lca-quorum for information on how the value is interpreted.


Use the pre-1.6.0 implementation for composing the alignment family. Requires --fs-engine = pt-server.

Search & Classify Options

--search-port=port_or_socket (:/tmp/sina_pt2_<pid>)

See --ptport. This option sets the port for the search database. It is only used if --search-db is specified and its value differs from the one given by --db.


Calculate the similarity of the query sequences with all reference sequences. Normally, SINA will only calculate the similarity for the sequences returned by a k-mer based similarity search. See also --search-kmer-candidates.

--search-no-fast don't use fast family search

See --fs-kmer-no-fast. This option configures the same behavior for the search stage.

--search-kmer-candidates=n (1000)

Configures the number of candidate reference sequences retrieved from the k-mer based search. For each candidate, the MSA based similarity is calculated and the search result based on these numbers. A value for n one or two orders larger than --search-max-result is usually quite sufficient.

--search-kmer-len=n (10)

See --fs-kmer-len. Sets k for the kmer based candidate search.

--search-kmer-mm arg

See --fs-kmer-mm. Sets the number of allowed mismatches within each kmer. Only available with the pt-server search engine.


See --fs-kmer-norel. Configures the candidate search to use absolute rather than length-relative scores for ordering the results.


Omit reference sequences of which the query is an exact sub-string from the result. Useful for testing and evaluation of the classification feature.


Specifies a (colon or comma separated) list of meta-data fields to be copied from each search result sequence into the output sequence. In the output sequence, the field names will each be prefixed with copy_<acc>_ where <acc> is the value of the acc field in the reference.

Use --arb-list-fields for listing the fields available in a given ARB database.

--search-iupac=[pessimistic|*optimistic|exact] (optimistic)

Configures how ambiguous bases are matched when computing the scores for the search results.

Ambiguous bases do not match anything because they could always be a mismatch.
Ambiguous bases are considered matches if a match with the other (potentially also ambiguous base) is possible. That is, N will match everything, including Y.
Matches on character level. N matches exactly N.
--search-correction=[none|jc] (none)

Apply distance correction to search result scores.

Leave score unmodified.
Apply Jukes-Cantor correction.
--search-cover=[abs|query|target|min|max|avg|overlap|all|nogap] (query)

Compute sequence similarity as the fraction of the number of matches and

the number 1: yields the absolute number of matching bases
the length of the query sequence. Yields the fraction of the query covered by the reference sequence.
the length of the target sequence. Yields the fraction of the result sequence covered by the query sequence.
the length of the shorter of the sequences compared.
the length of the longer of the sequences compared.
the average length of the two sequences compared.
the number of columns in which both sequences have bases. Yields the equivalent of matches / (matches+mismatches).
the number of columns in which either sequence has a bases. Similar to nogap, but does not ignore indel events.
the length of the overlapping portion of the two sequences.

Ignore lowercase bases when scoring result sequences. This can be used in conjunction with --lowercase=unaligned to ignore unaligned bases during the search and classification stage.