streme [options] --p <primary sequences>



<primary sequences>

The name of a file containing the primary (positive) sequences in FASTA format. The file must contain at least two valid sequences or STREME will reject it. Note that the command-line version of STREME does not attempt to detect the alphabet from the primary sequences, so you should specify it with the --dna, --rna, --protein or --alph options.


STREME writes its output to files in a directory named streme_out, which it creates if necessary. You can change the output directory using the --o or --oc options. The directory will contain:

STREME gives each motif it discovers a name, which is a consensus sequence that approximately describes the motif, prefixed by a number from 1 to the number of motifs found, e.g., "1-CCACYAGT". Motif names are intended to be mnemonic only, and are not intended to be used for searching sequences for matches; to search for matches to STREME motifs, use a STREME output file as input to a motif scanning program such as FIMO and MAST.


Option Parameter Description Default Behavior
--text Output in text format only to standard output. The program behaves as if --oc streme_out had been specified.
Objective Function
--objfun de|​ cd This option is used to select the objective function that STREME optimizes in searching for motifs.
deDifferential Enrichment This objective function scores motifs based on the enrichment of their sites in the primary sequences compared with the control sequences. STREME estimates motif enrichment using Fisher's exact test if the primary and control sequences have the same average length (within 0.01%), otherwise it uses the Binomial test.
cdCentral Distance This objective function scores motifs based on their tendency to occur near the center of the primary sequences, which must all be of the same length. No set of control sequences is allowed, and the primary sequences should include adequate flanking region around the expected motif sites—e.g., use sequences of 500bp for ChIP-seq. STREME estimates the tendency of a motif to occur near the centers of primary sequences using the cumulative Bates distribution applied to the mean distance of the best site from the sequence center.
STREME uses the Differential Enrichment (de) function.
Control Sequences and Hold-out Set
--n control sequences The name of a file containing control (negative) sequences in FASTA format. The control sequences must be in the same sequence alphabet as the primary sequences. If the average length of the control sequences is longer than that of the positive sequences, STREME trims the control sequences so that both sets have the same average length. If you do not provide control sequences, STREME creates them by shuffling a copy of each primary sequence, preserving the frequencies of words of length k (see next option). Shuffling also preserves the positions of non-core (e.g., ambiguous) characters in each sequence to avoid artifacts.
--kmer k Preserve the frequencies of words (k-mers) of this size when shuffling primary sequences to create control sequences. k must be in the range [1,..,6]. STREME also estimates a background model of order k-1 from the primary (positive) sequences for use in log-likelihood scoring of motif sites. STREME preserves the frequencies of words of length 3 (DNA and RNA), and 1 (Protein and Custom alphabets), and constructs background models of order 2 (DNA and RNA), and order 0 (Protein and Custom).
--hofract hofract The fraction of the primary sequences that STREME will randomly select and hold out for accurately estimating the significance of motifs.
Note: If the hold-out set would contain fewer than 5 sequences, STREME does not create it, and the motif p-values will be inaccurate.
Note: The letter frequencies in the final motifs reported by STREME are based on all primary sequences, including those in the hold-out set.
STREME reserves 0.1 (10%) of the primary sequences for estimating the significance of motifs.
--totallength len Restrict the maximum total length of the sequences used by STREME from the sequence file(s) to a total of at most len. The input sequences are first sorted alphabetically by sequence content, and then their order is randomized. (Potential) sequences are then assigned to the primary and hold-out sets. Then sequences are added in to the primary set in a the random order, skipping any potential sequence that would cause the total length of the primary set to exceed len times (1 - hofract). This is repeated for the hold-out set, not adding a sequence if the total length of the hold-out set would exceed len times hofract. The total length of the input sequences is not limited.
--seed seed Random seed for shuffling and sampling the hold-out set sequences (see above). STREME uses a random seed of 0.
Number of Motifs
--pvtpvt STREME will stop searching for motifs when too many consecutive motifs have p-values greater than pvt. (See option --patience, below.) 0.05
--patiencepatience Stop searching for motifs when patience consecutive non-significant motifs have been found. (Can be overridden by the --nmotifs or --time options; see below.) 3
--nmotifsnmotifs STREME will stop searching for motifs after finding nmotifs motifs. STREME stops when too many consecutive, non-significant motifs have been found.
--timemaxt STREME will stop searching for motifs if it has found at least one motif and it estimates that finding any more will cause the total running time to exceed maxt CPU seconds. STREME stops when too many consecutive, non-significant motifs have been found.
Motif Width
--minwminw Search for motifs with a width ≥ minw. Searches for motifs with a minimum width of 8.
--maxwmaxw Search for motifs with a width ≤ maxw. Searches for motifs with a maximum width of 15.
--ww Search for motifs with a width of w. Overrides --minw and --maxw. See --minw and --maxw, above.
Seed Evaluation and Refinement
--nevalneval The number of seed words of each width from 3 to maxw to evaluate for enrichment of approximate matches. 20
--nrefnref The number of seed words of each width from minw to maxw to convert to motifs and optimize using an iterative refinement algorithm. 4
--niterniter The iterative refinement algorithm is run for niter iterations, or until convergence, which ever comes first. 20
--desc description Include the text description in the HTML output. No description in the HTML output.
--dfile dfile Include the first 500 characters of text from the file named dfile in the HTML output. This option overrides option --desc. No description in the HTML output.
--verbosity1|2|3|4|5 A number that regulates the verbosity level of the output information messages. If set to 1 (quiet) then STREME will only output error messages, whereas the other extreme 5 (dump) outputs lots of information intended for debugging. The verbosity level is set to 2 (normal).

STREME algorithm overview

STREME searches for motifs by iterating the following five steps to until the selected stopping criterion is met. The stopping criteria are described in the "Number of Motifs" section, above.

  1. Suffix Tree Creation.

    STREME builds a single suffix tree that includes both the primary and control sequences (but not the hold-out set sequences).

  2. Seed Word Evaluation.

    STREME uses the tree to efficiently evaluate all words of length up the maximum motif width that occur in the primary sequences, computing the p-value of each such word's relative enrichment in the primary sequences using the chosen objective function. (Note: With the Differential Enrichment objective function, STREME will use the Binomial test instead of Fisher's exact test if the primary and control sequences have different average lengths. With the Central Distance objective function, STREME computes the cumulative Bates distribution of the average distances of the seed word from the centers of the sequences.)

  3. Motif Refinement.

    STREME converts each of the best seed words into a motif, and iteratively refines each motif, selecting the motif that best discriminates the primary sequences from the control sequences. At each iteration of refinement, the current motif and the (k-1)-order background are used with the suffix tree to efficiently find the best site in each sequence. The primary and control sequences are then sorted by the log-likelihood score of their best site, and the score threshold that optimizes the p-value of the statistical test (which depends on the chosen objective function) is found. The iteration ends by estimating a new version of the motif from the single best site in each primary sequence whose score is above the optimal threshold. This new motif is used in the next refinement iteration. Refinement stops when the p-value fails to improve or the maximum allowed number of iterations (20) have been performed.

  4. Motif Significance Computation.

    STREME computes the unbiased statistical significance of the of the motif by using the motif and the optimal discriminative score threshold (based on the primary and control sequences) to classify the hold-out set sequences, and then applying the statistical test (Fisher's exact test, Binomial test, or the cumulative Bates distribution) to the classification. Classification is based on the best match to the motif in each sequence (on either strand when the alphabet is complementable).

  5. Motif Erasing.

    STREME "erases" each of the sites of the best motif from both the primary and control sequences by converting the sites to the separator character. Only the positions in the site where the letter has a positive likelihood ratio are erased to allow some overlap of sites of different motifs.

STREME Running time and memory usage

The running time and memory usage of STREME depends on the total size of the sequences in its input, the length of the sequences, the alphabet of the sequences and the minimum and maximum motif widths. The following tables show the running times and memory usage for STREME on on different size datasets (of random sequences), as a function of the length of the sequences, on a 3.2 GHz Intel Core i7 processor with 16GB of memory. STREME was run using a single thread, and the motif width was the maximum allowed range: minimum = 5, maximum = 30. The total sizes of the sequences in the datasets from 100,000 (1 x 105) to 20,000,000 (2 x 107).

With very short sequences (length = 5), STREME is extremely fast, processing 4,000,000 DNA sequences in less than 100 seconds. The running time increases with sequence length, reaching a maximum for sequences around 30 long. With very long sequences, STREME is runs more quickly, with length 10,000 sequences taking approximately the same time per motif discovered as length 10 sequences. STREME runs about twice as slow on DNA sequences as on RNA sequences because STREME treats DNA sequences as double-stranded and RNA sequences as single-stranded. Because of the larger alphabet size, STREME runs about 5 times more slowly on protein sequences than on RNA sequences, which is the same factor as the ratio of the alphabet sizes (20/4). The running time of STREME with sequences over a custom alphabet will similarly be (roughly) proportional to the number of letters in the custom alphabet.

STREME Running Time (seconds / motif)
DNA RNA Protein
STREME Virtual Memory Size (bytes)
DNA RNA Protein


If you use STREME in your research, please cite the following paper:
"STREME: Accurate and versatile sequence motif discovery", bioRxiv, preprint, Nov. 23, 2020. [full text]