[an error occurred while processing this directive]
On this page:

Notes on Using Fluxus

The proper name is "Network" by Fluxus Engineering. But, most people just call it "Fluxus". This page is to convey some observations and tips for using it in branch analysis.

The software is designed to handle many types of taxonomic, genetic and linguistic information. We'll discuss it primarily from the standpoint of yDNA STR data.

Diagrams from Network by Fluxus Engineering.


We've found this program to be very helpful in analyzing ySTR data. The primary advantage of this analytical technique is that it yields a pictorial representation of numerical (or binary) data. A picture can be worth -- not only a thousand words -- but also many rows and columns of numbers.

Fluxus isn't the only program that will make phylogenetic diagrams or trees. But, unlike most others, it's free. The publisher requires only that one gives proper attribution when using its output. (Another free program is PopArt.)

It is available from Fluxus Engineering; just download and unzip. Be sure to download the User's Guide too; you'll need it. We advise creating a shortcut on your desktop with the Fluxus icon.

The software uses differences (such as STR marker/allele values) between taxa (group members) to infer or hypothesize how they're genetically related. It then displays the hypotheses visually and can save the display to a graphics format for later use.

Other types of data (SNPs, linguistic, anatomical) may also be used to construct network diagrams.


Things, in addition to the usual problems of interpreting ySTR data, to bear in mind:

  1. The diagrams are not proof; they visualize inferred, hypothesized genetic relationships. They are no better evidence than the input data. (ySTR data is "noisy"; it often is influenced by unknown and random factors.)
  2. It is critical that -- before applying the method -- the user establish reasonable confidence that the individuals are indeed genetically related. Failure to do so will result in a messy and meaningless diagram.
  3. The diagrams produced will be two-dimensional (duh!); the actual network may exist in three or more dimensions.
  4. The software uses the rule of maximum parsimony (Occam's Razor) to draw the shortest possible paths between nodes. This rule may not apply to all cases. Alternatives to maximum parsimony include maximum likelihood and Bayesian inference.
  5. Back-mutations are treated as no mutation; this can disguise the true network topology. .
  6. The User's Guide warns against using DYS389i or DYS389ii. Though reported by FTDNA & other labs as two markers, they actually consist of four markers combined into two for reporting.
  7. Multi-copy marker values (DYS385, DYS464, etc.) are not reported correctly, as their positions can not be identified.
  8. In general, the bigger the group, the more complex the diagram.
  9. There is an ISOGG page, http://www.isogg.org/phylo.htm, which gives instructions. Written in 2005 & 2006, it contains some errors as to later versions of the software (now with a 2016 update).

Getting Started

Download and read through the User's Guide, Much of it won't make sense at this early stage but you'll gain an idea of where to look up things.

Look at the work flow diagram on the User's Guide page 5; it outlines the steps to follow.


A branching diagram showing the cladistic relationships between a number of species of individuals, e.g., a "family tree". Fluxus network diagrams are a type of cladogram.
A method of classification of animals and plants according to the proportion of measurable characteristics that they have in common. It is assumed that the higher the proportion of characteristics that two organisms share, the more recently they diverged from a common ancestor.
Taxon, taxa
A taxon (plural taxa) is a person's (or other organism's) data. For human yDNA STR, the data are the tested allele values of the markers. Taxa may be identified by up to 9 alphanumeric characters and are represented by circles.
Locus, loci
A locus (plural, loci) is -- in this context -- a ySTR marker. Loci may be identified by up to 9 alphanumeric characters.
A node is a place on the network. It may consist of one taxon or many identical taxa. A multi-taxa node is identified by the ID of the first taxa in your data set to have those same characteristics. A node can also be either a modal haplotype or a "median vector" calculated by Fluxus.
Median vector
A median vector is a hypothesized node, inferred from the network data to create the shortest path. It may represent either a missing ancestral haplotype or alternative paths between real taxa. Median vectors are represented bu diamonds and named in the order created by Fluxus, "mv1", "mv2", etc.
A link is the represented path between two nodes. Directionality is not indicated but may be inferred, e.g., outward from the root. Links are represented by lines.
A torso is a network within the network. It typically consists of alternative paths between nodes.
A root is the presumed ancestor to all taxa in the group, perhaps the modal haplotype for the group or for the haplogroup. (A haplogroup modal could be considered an "outgroup" to provide a root.) A specific root may be included or not.
Maximum parsimony
Maximum parsimony ("Occam's Razor") means to make the fewest possible assumptions. The rule is that, when more than one possible explanation exists, the explanation requiring the fewest assumptions is usually the correct one.
Alternatives to the maximum parsimony rule include maximum likelihood and Bayesian inference. However, Fluxus does not support these criteria; they must be applied outside the Fluxus program; Fluxus uses only maximum parsimony.
Weights, weighting
A weight -- in the range of 0 to 99 -- is an index for likelihood; higher weights indicate less likelihood; lower weights more. Markers may be weighted in accord with average mutation frequencies. The default weight is 10; a weight of 0 causes a locus to have no effect. Notes:
The epsilon (ε) factor -- ranging from 0 to 231 -- relates to "connection cost" in a network calculated with median-joining (MJ) and works in conjunction with weights. Low numbers yield sparse networks and big numbers more complete networks. (Few networks are large enough for epsilon to be needed.)


There are at least three stages to go through in order to produce a diagram; each stage is indicated by a link on the program's top menu bar: These stage are

1. Data entry

This option is primarily for manual creation or editing of the input files. Due to user interface difficulty, we suggest manual data entry only for small networks with few taxa and loci to be entered. Or, it may also be used to make edits to previous input files.
  1. From the data entry drop-down menu, select either "Manual" (new file) or "Import rdf file".
  2. On the next menu, select Y-STR data; if importing, choose a .ych file.
  3. Click the Continue button. If manually entering, choose numbers of taxa and loci for input.
  4. In the next window, taxa are represented by rows and loci by columns. Default taxa names are "SEQ1", SEQ2, etc.; default loci names are "CH1", "CH2", etc.; marker values are at the intersections of rows & columns, with the default value of 99 (unknown); these may be edited.
  5. Save the file frequently while editing.
Alternative: Create the .ych file externally to Fluxus. Use McGee Utilities or a text editor such as Notepad. Fluxus, though, is picky about the file's format; the wrong format produces error codes and the program stops. The correct .ych file format is, by line number:
  1. The marker names, separated by commas
  2. The weights of the markers, separated by commas; can be left blank
  3. Blank line
  4. Taxon ID
  5. Marker values, separated by commas
  6. Frequency, i.e., number of taxa
  7. to 9, lines 10-12, 13-15, etc. -- repeat 4, 5, 6 (Each taxon requires 3 lines.)

2. Calculate Network

Once you have a .ych file, you can begin to calculate a network.
  1. Click on the Calculate link and select Network Calculations →  either Reduced Median or Median Joining
    1. Reduced Median yields a simplified network, useful to reduce complexity.
    2. Median Joining is recommended for the first run-through.
  2. A new window will open; in it open your .ych file. Menu sequence: File → Open. You may see a warning: "This file contains at least one N. The more Ns a file contains, the less reliable the phylogenetic analysis will be." ("N" refers to number of taxa; a minimum of 3 is needed to have any meaning.)
  3. "Parameters" on the menu gives you the option to change marker weights, epsilon, Frequency>1 criterion (active or inactive), Criterion Connection Cost, External rooting active or inactive), and MJ (median joining) square option active or inactive). Look them up in the User's Guide to see their effects.
  4. Calculate Network will carry out your calculations and create a .rdf file. It will then ask you to save as a .out file; do it.
  5. You will then see a message that the file has been saved successfully and you may proceed to Draw Network.

3. Draw Network

Once you have a .out file, you can begin to draw the network diagram. Click the Draw Network link on the top menu.
  1. Open your .out file. (Fluxus looks for a .sto file, but you don't have one and .out will do fine.)
  2. Aside: On my system, the Fluxus windows are larger than the monitor's screen. I have to resize them in order to access the controls.
  3. A partial network will be drawn; look for the "Continue" button to the lower right and click it. After a short time, a "Finalize" button will appear in the same place; click it too.
  4. You may now zoom in or out, changing the diagram's size and move it left, right, up or down in its window.
  5. You may also now change the diagram's appearance. You can increase font size from the default 6 points, minimum node size to 10 and maximum node size to 100.
  6. You may also move nodes in order to better separate them (and mutations; click and drag a node to its new place You may receive a warning that the link between nodes is not shown in proper length. (It is helpful to have links in rough proportion to number of mutations.)
  7. For multi-taxa nodes, I like to change colors of the pie slices to indicate the number of people in the node. I like one color per person and reserve magenta for the modal.
    1. Double-click on a node to see the taxa included.
    2. Right-click on a node to change the number of slices and their colors.
  8. Link colors and breadth may be changed; right-click on a link. I like 2 points in black for links outside the network torso and 4 points in green for links inside the torso.
  9. Colors and fonts of identifying labels may be changed; right-click on a label.
  10. Save your work (frequently) as a .fdi (formatted diagram) file.
  11. When finished, save as a bitmap (.bmp) image. (Be sure to change the filename extension in the save dialogue box.)
Because bitmaps are bulky, I like to edit the .bmp file in a graphics software program (Paint, Inkscape, etc.) and save as a .png or .jpg file. Paint permits cropping the image to eliminate wasted space. Inkscape permits adding a label.  

4. Time (TMRCA) Estimates

Click the top menu link and open a .fdi  or .dia file. 
  1. Select Mutation rate (1 mutation every x years). the default is 20,180 years -- a bit much for ySTR. CDY, for example, mutates about once in 28 generations (850 years) on average.
  2. Click Specify ancestral node and click on a node you judge to be "ancestral".
  3. Click on Specify descendant nodes and click one or more nodes you judge to be descendants of the ancestral nodes.
  4. Click on Calculate time.
  5. Output (most likely estimates) will be age in mutations and age in years, and standard deviations both in mutations and years. (My experience is that the standard deviations are large relative to the estimates, probably as a result of small N.)

Additional menu items are

Network Drawing Adjustments

Fluxus allows quite a bit of diagram adjustment and manipulation, enabling esthetic and visual clarity enhancements. IMHO, visual clarity is vital; it is the purpose of the exercise.

You can move nodes. I like to put the modal node at the center of the diagram. This implies a radiating pattern, in which branches spread out from a central point and allows greater separation of nodes.

Putting the modal node at one edge (left, top or bottom) would imply some nodes are older than others. This may or may not be the case, as all taxa nodes represent currently or recently living people. (Median vectors may represent deceased people.)

I also like to arrange things (nodes, links and labels) so they aren't on top of each other. Click and drag on a node to move it. Mutation labels will rearrange to be evenly spaced along a link.

Node sizes

I like to have nodes proportional to the number of taxa in it. To achieve this, reset both the minimum and maximum node sizes in the controls to the right of the diagram.

Slice colors

A node may represent more than one taxon. I like to indicate this by having as many slices as taxa and giving each a contrasting color. To see the taxa included, double-click a node. To change number of slices and colors, right-click a node. (You can not have more slices than taxa included.)

What does the diagram tell you?

A Fluxus Network diagram is a cladogram, a visual representation of inferred relationships within the group. They infer who's more closely and more distantly related and, to a limited degree, how.

It is not "proof", or even evidence, any more than the data going into it. For various reasons, the inferences may be wrong but they are, IMHO, as close approximations as one can get with ySTR data. These inferences do not trump documented genealogy, but can serve as guides to future research.

I happen to be a fan of visualization. A picture can be worth a thousand words and, as well, a thousand calculated numbers.

Two more sample diagrams

Group I1-001

Group R1b-007

The above diagrams (thank you Fluxus Engineering) are based on 111-marker results, but

For a collection of more diagrams, see this page.

What can we learn from these diagrams?

Some characteristics are the same for both groups.

Some characteristics are different.

Files Used

Fluxus takes you through a series of files in its process, all with different filename extensions. They are (except for bitmaps & PDFs) plain-text files.

  1. .ych

    contains the initial data. Example for G-044:
    Comment: 37 markers have been entered but most aren't needed.
  2. .rdf

    contains results of network calculations
    Notice that the .RDF file is much smaller, due to data reduction and conversion to binary (presence/absence). Only two markers show differences and are carried over from the .ych file
    Two differences (mutations) have been found: DYS458aa & DYS449aa. 
  3. .out

    Derived from network calculation and used as input to network drawing
    Results of file C:\Documents and Settings\Ralph\My Documents\GENEALOG\Taylor\FTDNA-admin\Y-Results\Fluxus\Fl-G-44-37.rdf with reduction threshold= 2.00000000000000E+0000
    Equivalent taxa : 
    Tax.      1 with tax.      2
    Tax.      1 with tax.      3
    Final links : 
    Link    1 0 Seq.      1 with Seq.      4, mutations at 2        
    Link    2 0 Seq.      1 with Seq.      5, mutations at 1        
    Frequ. of taxon        1 :   1
    Frequ. of taxon        2 :   1
    Frequ. of taxon        3 :   1
    Frequ. of taxon        4 :   1
    Frequ. of taxon        5 :   1
    Weight of character      1 :  10
    Weight of character      2 :  10
    Mapping of character;     1;D458aa
    Mapping of character;     2;D449aa
    Mapping of taxon;     1;modal
    Other information of taxon;     1;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     2;16678
    Other information of taxon;     2;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     3;56040
    Other information of taxon;     3;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     4;13707
    Other information of taxon;     4;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     5;94522
    Other information of taxon;     5;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;

  4. .fdi

    describes the formatted diagram with user enhancements  (Some of the syntax seems similar to SVG protocol. )

  5. .out

    input to the diagramming function. G-044 example
    Results of file C:\..\..\..\Taylor\..\Fluxus\Fl-G-44-37.rdf with reduction threshold= 2.00000000000000E+0000
    Equivalent taxa : 
    Tax.      1 with tax.      2
    Tax.      1 with tax.      3
    Final links : 
    Link    1 0 Seq.      1 with Seq.      4, mutations at 2        
    Link    2 0 Seq.      1 with Seq.      5, mutations at 1        **************************************************
    Frequ. of taxon        1 :   1
    Frequ. of taxon        2 :   1
    Frequ. of taxon        3 :   1
    Frequ. of taxon        4 :   1
    Frequ. of taxon        5 :   1
    Weight of character      1 :  10
    Weight of character      2 :  10
    Mapping of character;     1;D458aa
    Mapping of character;     2;D449aa
    Mapping of taxon;     1;modal
    Other information of taxon;     1;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     2;16678
    Other information of taxon;     2;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     3;560400
    Other information of taxon;     3;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     4;13707
    Other information of taxon;     4;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Mapping of taxon;     5;94522
    Other information of taxon;     5;Phenotype;;Geography;;Lineage;;Group1;;Group2;;Group3;;
    Comment: Describing a picture in words takes many words.
  6. .bmp or .pdf

    image files can be used in other applications.