RAREFAN (RAyt/REpin Finder and ANalyzer) Manual


Using the online tool

Here you can upload the genomes that are to be analyzed for their REPIN content. The genomes should be in fasta format and only have the specified file extensions.


You can also upload RAYT query sequences. These should be fasta files containing aminoacid sequences and should end in “.fa” or “.faa”. If these are not provided then you can select a RAYT gene from P. fluorescens SBW25 (Group 3 RAYT (1)) or one from E. coli (Group 2 RAYT (1)) provided by the server on the next screen.


Finally, it is possible to upload a phylogenetic tree in Newick format for your fasta sequences. For the webserver to be able to use the provided tree the leaves of the tree have to match the sequence file names. If the tree is not provided then a tree will be reconstructed using andi (2).



(1)     Please select a reference sequence from the sequence files you submitted. The reference sequence will be used to identify a maximum of eight seed sequences. The seed sequences have to occur at least n times (specified in (4)) and are of length N (specified in (5)) and are the basis for determining sequence groups. The way sequence groups are created is described in detail in publication (3), in the section “Grouping of highly abundant oligonucleotides in SBW25”. If sequence groups are part of REPINs, then each sequence group can be used to define a REPIN type.


(2)     There are two very divergent RAYT groups (E. coli  group and P. fluorescens group). If you do not know what kind of RAYT is present in your genome then you may have to run RAREFAN twice, once with the SBW25 and once with the E. coli RAYT. If you have supplied a RAYT protein yourself you should be able to select it here. These sequences are then used as query proteins in a TBLASTN (4) search of the provided genomes. All sequences that are identified below the e-value threshold set in (8).


(3)     If you have provided a tree file in the previous step then you can select it here.


(4)     This parameter is required for identifying REPINs. Only sequences that occur more frequently than this value will be considered when identifying REPINs.


(5)     The seed sequence length for identifying REPINs.


(6)     If two seed sequences occur at a distance of less than this value, are considered to be part of the same sequence group. Larger values will lead to smaller sequence groups and smaller values will lead to larger sequence groups.


(7)     If a member of a REP sequence group is closer to a given RAYT than specified by this number then the REP group is linked to the RAYT gene (they will get the same color in the RAYT and REPIN plots). Smaller values will lead to fewer REPIN-RAYT associations and larger values will lead to more associations.


(8)     This e-value determines, which genes are identified as RAYTs in the genome.


(9)     If this box is ticked then REPINs are identified. REPINs are defined as sequences that consist of two seed sequences in inverted orientation that are found at a distance of less than 130 bp. If the box is unticked then the seed sequences (REPs) identified as described in (1) are used for all further analyses. This can be useful if REPINs are asymmetric (i.e. there is for example a deletion/insertion in either the 5’ or 3’ REP sequence). This is for example the case for E. coli REPINs.


(10) If you provide an email address here you will be notified once the job is finished. This is particular useful when applying RAREFAN to large datasets.



Once the job is submitted, it is assigned a unique identifier. If you have not supplied an email address you will either need to keep the site open or remember the unique ID to access the website later. If you have closed the website but know the unique ID you can access your data by calling this address http://rarefan.evolbio.mpg.de/results?run_id= and inserting your ID.



Once the job is finished you can access your data either by directly plotting the data (description of the plots below) (1); or by browsing through the folder structure online (2); or by downloading the data and viewing it on your hard disk (3). You can also rerun the data with a different reference or different parameters (4).


File output

All output data is located in a folder called out/.

The output files in the folder out/ include the following:

Files in out/



A phylogenetic tree of all genomes generated with andi (http://github.com/evolbioinf/andi/) and clustDist (http://guanine.evolbio.mpg.de/problemsBook/node1.html).



A file containing the frequencies of all 21bp long sequences found in the designated reference genome.




Contains all 21bp long sequences that occur more frequently than n (default 55) times in the reference genome. 


A file containing the nucleotide sequences of all RAYT relatives identified with BLAST+ in all provided genomes. Contains only sequences that are longer than 240 bp.


If REPINs of that group were identified in the strain and the REPIN option is ticked then it contains the most frequent REPIN identified for each sequence type in each strain. If the REPIN option is not ticked then the most common REP sequence is shown.


If the REPINs option is ticked during submission then it contains information on RAYT and REPIN numbers are shown as long as the strain does contain at least one copy of that REPIN type.

If the REPIN option is not ticked then only information on REP sequences are going to be available even if it says REPINs in the file.

Specifically, information on the number of RAYTs, the number of REPINs, the master sequence, the number of master sequences, the entire REP/REPIN population size, the number of REPIN clusters that contain more than 10 sequences, all REPINs in the population as well as all REPINs that differ to the master sequences in at most three nucleotides.


Nucleotide alignment of all RAYTs identified in each of the query genomes.


Phylogenetic tree calculated with PHYML from the above RAYT alignment.

rayt_[strain name].tab

Contains location information for each identified RAYT relative for each strain. The files can be viewed with artemis (5).


Contains for each strain the frequency of the six identified 21bp long seeds.


Table containing information on which RAYT (column 2) from which genome (column 1) is associated with which REPIN group(s) (column 3).


Same information as above but in a different format.


Nucleotide sequences of each RAYT gene from each of the genomes in FASTA format.


There is a subfolder called groupSeedSequences/. All 21bp long sequences in the genome that occur more frequently than 55 times are sorted into 6 sequence groups. These sequence groups are stored in the following files:

Files in out/groupSeedSequences/



Contains all seed sequences that occur more than 55 [default] times in the genome that were sorted into the specific group. As well as the frequency of those sequences in the reference genome. The most common sequence in each group is used as a seed sequence to determine REPIN populations across all submitted sequence files.


The same information as above but in FASTA format.



Contains the locations of all overrepresented 21bp long sequences in the reference genome. This file can be viewed in artemis (https://www.sanger.ac.uk/tool/artemis/) together with the reference genome file.


For each genome there are six output folders (ending in _0 to _5), corresponding to each of the six sequence groups (only if there are at least six sequence groups that occur more frequently than the seed sequence frequency threshold).

Each folder contains the following files:

Files in out/[genome]_[0-5]/



Degree distribution of the REPIN network, where each REPIN is a node. A REPIN is connected to another REPIN if they differ in exactly one position. The degree distribution is a histogram of the number of connections of all the nodes. 


For the largest sequence cluster determined by mcl (6) that consists of REPINs (two REPs in inverted orientation) this file contains the number of REPINs in each sequence class. Sequence class 0 is the master sequence. By definition the most common REPIN in the sequence population. Sequence class 1 contains all REPINs differing in exactly one position to the master sequence. Sequence class 2 contains REPINs differing in 2 positions etc.


Contains the clustering output by mcl. Each line contains the REPIN sequence IDs that belong to the same cluster. Lines are sorted by cluster size. REPIN/REP sequences are clustered on sequence similarity.


Contains the most common 21bp long sequence of this group and its frequency in the genome, which is the basis for identifying all related REP sequences and from those the REPINs formed by these REP sequences.


The identity and frequency of all REPINs and REP sequences.


The identity and frequency of all REPINs or REP sequences in the largest REPIN/REP sequence cluster.


Contains REPINs and REP sequences as well as their positions in FASTA format. Position information starts with the location in genome FASTA file (first sequence is 0...) followed by the start and end position of the entire REPIN/REP sequence.  


Same information as above just for the largest mcl cluster.

[genome]_[0-5]_[mcl cluster number].ss

The same information as above except that it contains the sequences for a specific cluster identified by mcl (see file *.mcl).


REP sequence information in FASTA format.


Location in tab format. Can be used to display locations of REPs and REPINs in the genome via artemis.

[genome]_[0-5]_[mcl cluster number].tab

Contains the location of REP/REPINs for each subcluster separately for viewing in artemis.


Contains network connections between nodes of all sequences. Can be used to view network in for example R or cytoscape together with the nodes file.


Contains information on which REPIN/REP cluster is in the proximity (within 200bp) of any of the RAYT genes identified in the genome.

subfolder [genome]_[0-5]/

Contains the complete sequences (including the variable region) for all identified REPs and REPINs.


Data Plots

Here you can select the REPIN/RAYT type that is being plotted. This is, for example, the data that is stored in the file presAbs_[number].txt.


The first plot shows the relationship between the RAYT genes and the REPIN type each RAYT gene is associated with. The tree was generated from a multiple sequence alignment of RAYT DNA sequences using the program MUSCLE (8). The tree itself is built using PHYML (9). The colors of the tip labels correspond to the associated REPIN types. Colors are usually monophyletic due to a strong association between RAYT and REPIN type.


The second plot shows the number of RAYTs and REPINs per genome. The tree on the left side of the figure shows the phylogeny of the submitted genomes. The tree was built applying neighbor joining (7) to a distance matrix generated with the program “andi” applied to whole genomes (2). The next column shows the presence and absence of the associated RAYT transposases. A RAYT transposase is considered associated when a REPIN of the type is found within 200bp of the transposase. The REPIN population size in each of the genomes is shown in the last column.





The proportion of master sequences (indicates sequence conservation) in a REPIN population and the REPIN population size. According to Quasispecies theory or mutation-selection balance (10, 11), the higher proportion of master sequences (the most common sequence in the population) correlates with higher duplication rates of the sequence population. The closer populations are to the lower left of the plot, the smaller and more decayed they are and the less likely they are to be alive (i.e. actively replicating). Only populations that are colored are associated with a RAYT transposase.


1.    F. Bertels, J. Gallie, P. B. Rainey, Identification and Characterization of Domesticated Bacterial Transposases. Genome Biol. Evol. 9, 2110–2121 (2017).

2.    B. Haubold, F. Klötzl, P. Pfaffelhuber, andi: fast and accurate estimation of evolutionary distances between closely related genomes. Bioinformatics 31, 1169–1175 (2015).

3.    F. Bertels, P. B. Rainey, Within-Genome Evolution of REPINs: a New Family of Miniature Mobile DNA in Bacteria. PLoS Genet. 7, e1002132 (2011).

4.    C. Camacho, et al., BLAST+: architecture and applications. BMC Bioinformatics 10, 421–9 (2009).

5.    K. Rutherford, et al., Artemis: sequence visualization and annotation. Bioinformatics 16, 944–945 (2000).

6.    A. J. Enright, S. Van Dongen, C. A. Ouzounis, An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res. 30, 1575–1584 (2002).

7.    N. Saitou, M. Nei, The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425 (1987).

8.    R. C. Edgar, MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).

9.    S. Guindon, et al., New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst. Biol. 59, 307–321 (2010).

10.  F. Bertels, C. S. Gokhale, A. Traulsen, Discovering Complete Quasispecies in Bacterial Genomes. Genetics 206, 2149–2157 (2017).

11.  F. Bertels, P. B. Rainey, “REPINs are facultative genomic symbionts of bacterial genomes” (2021).