The Dragon Phylogeny

Overview

A phylogeny is a tree that represents a hypothesis about the evolutionary relationships among organisms. Connected branches form a ‘clade’ and shorter connections represent more recent evolutionary divergence compared to long (i.e. deep) branches. There are many different methods for building a phylogenetic tree, but in general, trees are built using clustering algorithms that group objects by some measure of their similarity. Modern phylogenies are based on DNA or protein similarity, but in principal we can cluster objects based on any trait we can measure. To look at the process in more detail, let’s do a phylogeny of dragons.

The Dragon Phylogeny is a project developed when Rob Colautti was a postdoctoral researcher at the University of British Columbia. It was originally developed as a t-shirt Threadless.com, which is out of print now.

HOWEVER Some recent students have resurrected the project and re-released a slightly different version, which is available at https://dragonphylogeny.threadless.com/

100% of proceeds from t-shirt sales support grants & projects that diversity representation of STEM researchers, particularly in Ecology and Evolutionary Biology.

#diversifyEEB

The original design generated some media coverage:

• io9/Gizmodo Article – A nice write-up about the project on the blog io9.com (now part of Gizmodo).
• CBC.ca Interview – from an interview with CBC radio, later posted on their website.

GitHub

The project now lives on at GitHub: https://github.com/ColauttiLab/DragonPhylogeny

Scoring Traits

Images

The images are available as a pdf file on the Dragon Phylogeny GitHub site: Dragon_Pics.pdf NOTE This is a large file size (9.4MB). The PDF contains images of all dragons used in the original Dragon Phylogeny Unfortunately, we don’t have any dragon blood to do a DNA-based phylogeny. However, we can try to cluster dragons based on their physical appearance. The first step is to locate some pictures of dragons and choose a common set of traits.

Here are the traits for a variety of dragons:

Origin	RefNum	Teeth.Type	Tongue	Eyes	Eye.position	Ears	Nose	Snout	Horns	Skin.face.neck	Skin.body	Skin.Belly	Dorsal.ridge	Toes	Opposed.toes	Legs	Claws	Wings	Wing.type	Wing.structure	Body.Type	Tail.end	Size	wiskers
German	1	Blunt	forked; short	small	lateral	spearlike	upturned; forward nostrils	long	absent	smooth	fish scales	wide platemail	absent	3	0	4	short catlike	2	bat-like	4 ridges	elongate	point		absent
French	2	Fangs	forked; long	average	lateral	absent	upturned; forward nostrils	long	plates	hairy	soft scales	hair	absent	3	1	4	short talon	2	bird-like		elongate	point	1-2x human	absent
French	3	Blunt&Fangs	wormy; short		forward	spearlike	pointy; forward nostrils	long	absent	smooth	smooth	smooth	absent	3	0	4	short catlike	2	bat-like	11 ridges	rotund	point	1-2x human	absent
Dutch	4	Pointy		narrow	lateral	spearlike	upturned; forward nostrils	moderate	absent	scaly skin	round platemail	smooth	single row; spikes	3	1	4	short talon	2	bird-like	5major; 6minor ridges				absent
English	5			small	lateral	absent	upturned; lateral nostrils	long	absent	smooth	smooth	smooth	single row; continuous	2	0	4		2	batlike	5 ridges	elongate	point	1-2x human	absent
American	6	Pointy	single; long	large	lateral	absent	slits; forward	absent	absent	scaly skin	hexagonal scales	wide platemail	absent	0	0	0	0	0	0	0	snake		3x human	absent
French	7	Blunt&Fangs	forked; long	small	forward	spearlike	upturned; forward nostrils	blunt	absent	hairy	rough	wide platemail		4	1	4	large talon	2	mix	3major; 1 minor	elongate	point	1-2x human	absent
English	8	Pointy	long	average	lateral	small	slits; lateral	long; duckbill	short	fish scales	fish scales	fish scales	single row; jagged	3	0	2	short talon	2	batlike	7major; 3minor	elongate	split (Y shape)	1-2x human	absent
French	9	Pointy		average	forward	round	upturned; forward nostrils	blunt	absent	hairy	smooth	smooth	absent	3	1	4	short talon	2 wing-leg hybrid	batlike	6 ridges	rotund	point	1-2x human	absent
French	10	Pointy		average	forward	large dishlike	upturned; forward nostrils	blunt	absent	hairy	extra wide plates	see body	absent	3	0	2	short talon	2 wing-leg hybrid	batlike	7 ridges	elongate	point	1-2x human	absent
Spanish	11	Blunt&Fangs		small	lateral	round	hooked	long	absent	smooth		smooth	absent	3	1	4	short talon	2	batlike	4 ridges	elongate	point	1-2x human	absent
Japanese	12			average	forward	round	small; forward	long	long	hairy	fish scales	wide platemail	single; overlapping plates	3	1	4	large talongs	absent			snake	point		long
Japanese	13			narrow	forward	absent	upturned; forward nostrils	long	absent	spiny	spiny		absent	3	NA	NA					snake			short
Japanese	14	Fangs&Pointy		large	forward	absent	upturned; forward nostrils	long	short	hairy	fish scales	wide platemail	single		NA	0		0			snake	blunt	3x human	short
Japanese	15	Fangs&Pointy		large	lateral	absent	upturned; forward nostrils	blunt	jagged	spiny	fish scales	wide platemail	absent		NA	0		0			snake		1-2x human	long
Japanese	16	Pointy	short	average	lateral	absent	upturned; forward nostrils	long	2lobed antler	hairy	rough	wide platemail	single; jagged	3	0	2	long catlike	0			snake			long
Japanese	17	Pointy		average	forward	spearlike	upturned; forward nostrils	long	2lobed antler	spiny	fish scales	wide platemail	single; spiked	3	0	4	large talon	absent			snake	blunt		long
Japanese	18	Fangs&Pointy	short	large	forward	absent	upturned; forward nostrils	blunt	absent	spiny	large scales	wide platemail	absent		NA	4		0			snake		1-2x human
Japanese	19	Fangs&Pointy		large	forward	absent	upturned; forward nostrils	blunt	short	rough	fish scales	wide platemail	absent		NA	0		0			snake		1-2x human	long
Japanese	20			average	forward	round	upturned; forward nostrils	blunt	short	hairy	fish scales	wide platemail	single; jagged	3	0	4	short catlike	0			snake			long
Japanese	21	Fangs&Pointy	short	small	forward	absent	upturned; forward nostrils	long	jagged antlers	hairy	fish scales	wide platemail	single; spiked	3	0	4	large talon	0			snake			long
Japanese	22			large	forward	spearlike	upturned; forward nostrils	long	short antler	hairy	fish scales	wide platemail	single; spiked	3	0	4	large talon	0			snake	fan		medium
Japanese	23			average	forward	round	upturned; forward nostrils	long	2lobed antler	hairy	fish scales	wide platemail	single; jagged	3	0	4	large talon	0			snake	fan	1-2x human	medium
Japanese	24			large	lateral	spearlike	forward	long	2lobed antler	hairy	fish scales	wide platemail	single; spiked	3	0	4	large talon	0			snake	fan		medium
Japanese	25	Fangs&Pointy		small	forward	absent	upturned; forward nostrils	long	long	hairy	bumpy	double plates	single jagged	3	0	4	large talon	0			snake	fan		long
Japanese	26	Fangs&Pointy	short	average	forward	spearlike	upturned; forward nostrils	long	2lobed antler	hairy	bumpy	double plates	absent	3	0	NA	large talon	0			snake			medium
Japanese	27		short	small	forward		upturned; forward nostrils	long	long	hairy	fish scales	double plates	single; overlapping plates	3	0	4	large talon	0			snake	point	5-6x human	long
Japanese	28	Fangs&Pointy	short	small	forward	spearlike	upturned; forward nostrils	long	2lobed antler	hairy	fish scales	wide platemail	single; jagged	3	0	4	large talon	0			snake	fan	3-4x human	long
Japanese	29	Fangs&Pointy	short	average	forward	spearlike	upturned; forward nostrils	long	very long; multilobed	hairy	fish scales	double plates	single; jagged	4	0	4	large talon	0			snake	point	1-2x human	absent
Italian	30	Pointy		small	forward	round	pointy	long	absent	smooth	small scales	smooth	single; ridge	3	0	4	short talon	2	birdlike	6 ridges	elongate	point	1/2 human	absent
Italian	31	Pointy	short	average	forward	small	upturned; forward nostrils	moderate	absent	smooth	light hair	smooth	absent	3; webbed	0	4	short catlike	2	batlike	5 ridges	elongate	point	1/2 human	absent
Italian	32	Fangs	short	small	forward	spearlike	upturned; forward nostrils	moderate	absent	hairy	light hair	smooth	absent	3	0	2	short catlike	2	batlike	9major; 2 minor	elongate	point	1-2x human	absent
	33	Fangs&Pointy		average	forward	spearlike	small; forward	long	stubs	scaly skin; hair	fish scales	hair	single; spiked	4	0	4	short talon	2	batlike	5 ridges	elongate		1-2x human	absent
German	34	Pointy	double	small	lateral	absent	small; forward	long	absent	smooth	hairy	double muscular	single; hairy	3	1	2	short talon	2	batlike	2 ridges	elongate	point	1-2x human	absent
English	35	Pointy	forked; long	small	forward	spearlike	lateral	long	medium	skaly skin	fish scales	wide platemail	single; jagged	4	1	4	large talons	2	mix	5 ridges	elongate	point	1-2x human	nose horn (modified wisker?)
German	36	Fangs&Pointy	short	large	forward	spearlike	upturned; forward nostrils	long	stubs	hairy	light hair	light hair	sigle; ridge	4	1	4	short catlike	2	batlike	4 ridges	elongate		1-2x human	absent
Dutch	37	Fangs&Pointy	short	average	forward	spearlike	forward	moderate	absent	smooth	smooth	muscular	single bumps	3	0	2	large talons	2	batlike	3 ridges	elongate	point	1-2x human	absent
Spanish	38	Pointy	forked; long	small	forward	small	small; forward	long	absent	smooth	smooth	smooth	wide overlapping plates	3	1	2	short talon	2	batlike	5 ridges	elongate	point	1-2x human	absent
Italian	39	Pointy	long	small	forward	absent	pointy	forward	jagged antlers	rough	lizard scales	lizard scales	single bumps	3	1	4	short catlike	2	birdlike	5 ridges	elongate	point	1/2 human	absent
Italian	40	Pointy	long	small	forward	spearlike	forward	long	absent	rough	hairy	muscular	absent	4	1	4	short catlike	2	batlike	17+ ridges	elongate	point	1-2x human	absent
English	41	Pointy	narrow; short	average	lateral	small	forward	long	absent	hairy		lizard scales		5	0	4	short talon	2	batlike	5 ridges	elongate		1-2x human	absent
Italian	42		narrow; short	small	lateral	spearlike	forward	long	absent	rough	smooth	bump ridge	absent	4	1	4	short talon	2	batlike	5 ridges	elongate	point	1-2x human	absent
Spanish	43	Pointy	narrow; short	small	lateral	round	pointy	long	absent	hairy	complex leather	wide platemail	single; complex bumps	3	1	4	short talon	2	batlike	5 ridges	elongate	point	1-2x human	absent
Italian	44	Pointy	long	small	lateral	small	upturned; forward nostrils	long	absent	rough	bumpy				NA	4		2	batlike	9 ridges	elongate	point	1-2x human	absent
Italian	45	Fangs&Pointy	long	small	lateral	absent	pointy	long	absent	smooth	smooth	smooth	single; ridge	3	0	2	short catlike	2	batlike	9 ridges	elongate	point	1-2x human	absent
English	46	Pointy		narrow	lateral	absent	forward	long	stubs	smooth	platemail	smooth	absent	4	0	4	small talons	0				blunt	1-2x human	absent
Italian	47	Pointy	narrow; short	small	forward	spearlike	forward	long	fanlike	hairy	hairy	smooth	absent	3	1	4	small talons	2	batlike	7 ridges	elongate	point	1-2x human	absent
Dutch	48	Pointy	long	average	forward	spearlike	forward	moderate	stubs	smooth	light hair	smooth	single; bumped	3	1	4	small talons	2	birdlike	5 ridges	elongate	point	1/2 human	absent
Indian	49	Pointy	long	small	lateral	absent	forward	long	absent	rough	bumpy	serrated	single; ridge	3	1	4	absent	0				point		absent
Japanese	50			narrow	forward	small	upturned; forward nostrils	long	long	hairy	fish scales	muscular	single; overlapping plates	3	0	4	large talons	0			snake	point	1-2x human	absent
Japanese	51	Fangs		narrow	forward	spearlike	upturned; forward nostrils	long	long	hairy	rough	muscular	single; overlapping plates	3	0	4	small talons	2	birdlike	dozens ridges	snake	blunt	1-2x human	absent
Japanese	52			narrow	lateral	round	upturned; forward nostrils	long	absent	hairy	fish scales	muscular	absent	3	0	4	large talons	0			snake	point	1/4 human	medium
Japanese	53			large	forward	absent	forward	long	absent	hairy	lizard scales	lizard scales	wide overlapping plates	3	0	NA	small talons	0				point	1-2x human	small
Iranian	54				forward		forward	long	fanlike	smooth	smooth	smooth	single; ridge	3	0	4	short catlike	2	birdlike	7 ridges	snake		1-2x human	absent
Iranian	55	Fangs		small	lateral	small	forward	long	small	smooth	smooth	smooth	single; ridge	3	0	4	small talons	0			snake	point	1-2x human	absent
Iranian	56	Fangs	many points	narrow	lateral	spearlike	small; forward	long	long	rough	rough	rough	absent	4	1	2	small talons	2	batlike	4 ridges	snake	point	1-2x human	absent
Iranian	57	Fangs	narrow; short	small	forward	small	upturned; forward nostrils	long	small	rough	rough	muscular	absent	4	1	4	small talons	0			snake	point	1-2x human	absent
Turkish	58	Pointy	long	average	lateral	round	small; forward	long	small	rough	rough	muscular	single; ridge		NA	0		0				blunt	1-2x human	absent
Iranian	59	Pointy	long	small	forward	spearlike	small; forward	long	2lobed antler	rough	fish scales	fish scales	single; jagged	4	0	4	small talons	0			snake	point	1-2x human	absent
Iranian	60	Fangs	narrow; short	narrow	forward	absent	forward	long	absent	hairy	smooth	muscular	single; ridge	4	0	4	long catlike	0			snake	point	1-2x human	absent
Turkish	61	Pointy	forked; long	narrow	lateral	absent	forward	long	medium	rough	smooth	muscular	single; ridge	4	0	4	small talons	0			snake			absent
Turkish	62	Fangs&Pointy	forked; long	small	lateral	spearlike	forward	long	medium	smooth	smooth	muscular	single; ridge	4	0	4	small talons	0			snake	blunt		absent
Ukraine	63	Blunt&Pointy	short	small	lateral	round	upturned; forward nostrils	forward	ramhorn	rough	lizard scales	muscular	hair	3	0	4	short catlike	0			elongate	point	1-2x human	absent
Ukraine	64	Pointy		average	lateral	absent	tiny	long	short	smooth	smooth	rough	absent	4	0	2	long catlike	0			snake	threetail	1-2x human	absent
Russia	65	Pointy	long	average	lateral	absent	forward	moderate	absent	rough	rough	spiny	single; bumped	2	1	2	short catlike	2 wing-leg hybrid	birdlike	6 ridges	snake	blunt	1-2x human	absent
Ukraine	66	Pointy		small	lateral	round	lateral	moderate	absent	smooth	smooth	smooth	single; bumped	3	0	4	small talons	2	batlike	5 ridges	elongate	point	1-2x human	absent
Russia	67		short	narrow	lateral	absent	upturned; forward nostrils	moderate	small	hairy	hairy	hairy	absent		NA	2		2 wing-leg hybrid	birdlike	5 ridges	snake	point	1/2 human	absent
Greece	68	Fangs	short	average	forward	spearlike	forward	moderate	absent	hairy	hairy	hairy	single; bumped	3	1	4	large talons	2	birdlike	9 ridges	elongate	point	1-2x human	absent
Italian	69	Pointy	long	average	lateral	spearlike	upturned; lateral nostrils	long	absent	rough	smooth	smooth	single; spiked	4	NA	4	small talons	2	birdlike	6 ridges	elongate	point	1-2x human	absent
American	70			small	forward	large speaklike	upturned; forward nostrils	long	absent	rough	rough	wide platemail	absent	4	NA	4	small talons	absent				point	1-2x human	absent
British (Wales)	71		long; spearhead	narrow	lateral	spearlike	forward	long	absent	hairy	fish scales	lateral plates	single; spiked	3	1	4	large talons	2	batlike	9 ridges	elongate	spear		absent
British	72	Fangs	long; spearhead	narrow	lateral	large spearlike	tiny	long	absent	smooth	smooth	lateral plates	single; jagged	2	1	2	small talons	2	batlike	10 ridges	elongate	spear		absent
British	73	Fangs	short	narrow	lateral	spearlike	tiny	long	small	spiny	spiny	lateral plates	single; jagged	4	0	2	small talons	2	batlike	6 ridges	elongate	spear		absent
British	74	Fangs	long; spearhead	narrow	lateral	spearlike	forward	long	absent	rough	rough	lateral plates	single; jagged	4	0	4	small talons	2	batlike	5 ridges	elongate	spear		absent
British	75	Fangs	long; spearhead	average	lateral	spearlike	forward	long	absent	hairy	lizard scales	muscular	single; jagged	4	0	2	large talons	2	batlike	5 ridges	elongate	spear		absent

Encoding

Now that we’ve scored the traits, we have to encode them – in this case we’ll use 1s and 0s, with ? indicating unknown values for the traits that couldn’t be observed in some of the photos. The coding is easy for binary traits (present/absent), however most traits are not binary, and we might want coding that will account for inferred evolutionary transitions. For example, if we look at skin type we have several categories:

• fish scales
• spiny
• hairy
• plates
• scaly skin
• bumps/ridged skin
• smooth

What if we want to encode an evolutionary model? For example, one that looks like this:

spiny <-- fish scales --> scaly skin  
  |           |              /   \
  v           v             v     v
hairy       plates      smooth   bumpy

We need a coding system where the coding of the derived states is more similar to the ancenstral form than to the other states.

How would you code these using a binary vector?

Here’s one way:

100000 <-- 000000 --> 000100  
  |           |        /   \
  v           v       v     v
110000     001000  000110  000101

Trait coding

Here is the full list of traits and how they were encoded.

## [1] "Order"     "Trait"     "Phenotype" "Binary"

Order	Trait	Phenotype	Binary
1	Appendages	Zero	1111
1	Appendages	Two	1101
1	Appendages	Four	1001
1	Appendages	Six	0001
2	Mass	1/4 human	0000
2	Mass	1/2 human	0001
2	Mass	1-2x human	0011
2	Mass	3-4x human	0111
2	Mass	>4x human	1111
3	Body type	Rotund	00
3	Body type	Elongate	01
3	Body type	Snakelike	11
4	Claw type	Long Catlike	1100
4	Claw type	Short Catlike	1000
4	Claw type	Absent	0000
4	Claw type	Short Talons	0010
4	Claw type	Long Talons	0011
5	Dorsal ridges	Plates	100000
5	Dorsal ridges	Absent	000000
5	Dorsal ridges	Bumps	010000
5	Dorsal ridges	Spike	011000
5	Dorsal ridges	Ridge	010100
5	Dorsal ridges	Jagged	010010
5	Dorsal ridges	Hair	010011
6	Ear morphology	Absent	000
6	Ear morphology	Round or Small	100
6	Ear morphology	Spearlike	010
6	Ear morphology	Other	001
7	Eye morphology	Avg	000
7	Eye morphology	Large	001
7	Eye morphology	Narrow	010
7	Eye morphology	Small	100
8	Eye position	Lateral	0
8	Eye position	Forward	1
9	Horn type	Absent	000
9	Horn type	Stubs/Small	100
9	Horn type	Med/Long	110
9	Horn type	Jagged/Antlers	111
10	Nose Position	Lateral	0
10	Nose Position	Forward	1
11	Nasal morphology	Upturned	1
11	Nasal morphology	other	0
12	Skin-dorsal	Fish Scales	000000
12	Skin-dorsal	Spiny	000100
12	Skin-dorsal	Hairy	000110
12	Skin-dorsal	Plates	000001
12	Skin-dorsal	Scaly Skin	100000
12	Skin-dorsal	Rough Skin	101000
12	Skin-dorsal	Smooth Skin	110000
13	Skin-head	Fish Scales	000000
13	Skin-head	Spiny	000100
13	Skin-head	Hairy	000110
13	Skin-head	Plates	000001
13	Skin-head	Scaly Skin	100000
13	Skin-head	Rough Skin	101000
13	Skin-head	Smooth Skin	110000
14	Skin-ventral	Fish Scales	000000
14	Skin-ventral	Spiny	000100
14	Skin-ventral	Hairy	000110
14	Skin-ventral	Plates	000001
14	Skin-ventral	Scaly Skin	100000
14	Skin-ventral	Rough Skin	101000
14	Skin-ventral	Smooth Skin	110000
15	Snout type	Absent	0000
15	Snout type	Beak	0001
15	Snout type	Blunt	1000
15	Snout type	Moderate	1100
15	Snout type	Long	1110
16	Tail type	Blunt/Point	10
16	Tail type	Fan/Split Y	00
16	Tail type	Sepear	01
17	Teeth	Pointy Only	0000
17	Teeth	Blunt + Pointy	1000
17	Teeth	Blunt Only	1100
17	Teeth	Fangs + Other	0001
17	Teeth	Fangs Only	0011
18	Toes-opposing	Yes	0
18	Toes-opposing	No	1
19	Toe Number	> Five	000000
19	Toe Number	Five	100000
19	Toe Number	Four	110000
19	Toe Number	Three	111000
19	Toe Number	Two	111100
19	Toe Number	One	111110
19	Toe Number	Zero	111111
20	Tongue length	Short	0
20	Tongue length	Long	1
21	Tongue morphology	Regular	00
21	Tongue morphology	Forked	01
21	Tongue morphology	Spear	10
22	Ventral plates	Yes	1
22	Ventral plates	No	0
23	Whiskers	Absent	00
23	Whiskers	Short	10
23	Whiskers	Long	11
24	Wing structure	Absent	00
24	Wing structure	Hybrid	10
24	Wing structure	Full	11
25	Wing type	Bat	100
25	Wing type	Bird	010
25	Wing type	Hybrid	001

To encode a dragon, traits are first arranged by the Order column, and then the observed Phenotype for each Trait is recoded as 1s and 0s (or ? for missing values) using the corresponding Binary code. Finally, all of the 1s and zeros are combined into a single vector.

Nexus Data File

The file DragonMatrix.nex contains the encoded traits, along with a few extra lines of information that specify the data in a nexus data file format. Nexus files are just readable text files that follow a few formatting rules, typically with a .nxs or .nex file name extension. For more information, see the Wikipedia Entry or Christoph Champ wiki.

Open up the file in a text editor:

The penultimate line is a semicolon ;, which specifies the end of the coded characters, followed by End; indicating the end of the file.

The first few lines set up the data format:

#NEXUS
begin data;
dimensions ntax=77 nchar=78;
format datatype=binary interleave=no gap=?;
matrix

Here’s a breakdown of what these first few lines do:

• The first line shows us that it’s a #NEXUS data file
• begin data specifies the start of the data
• The dimensions line specifies the number of taxa (n=77 dragons) and characters (t=78 binary trait scores)
• The fifth line denotes that the data are arranged in a matrix format.
• The format line notes that the traits are encoded as binary. The gap specifies the symbol used for data gaps – i.e. missing values. The interleave=no specifies that each line contains all of the traits for the dragon. If this were DNA we might have 1000 or more base pairs in our sequence. In that case, we probably wouldn’t want a single line of base-pairs. Instead we might break it up into smaller chunks in an interleave format, like this:

Dragon1 TTGTCGAGTGTGCGGCAGCTTAGGTGAATTAAGTCCGGGCAACCTTTAGT
Dragon2 CAATAGCATACTACCGTGCGAGCCAGCTTATAGGTCGTTGCAGGTTATTA
Dragon3 ATGTCATTCGCCACGAGACTTTACTAGGGTATCATGCCGAAAGGGGATGG

Dragon1 TGTCCTGTGTGGGAAGTCGTGCCAGGACGGTTACAGCCTTAGCTTGTGCG
Dragon2 AAGCGAACTGAAGCGGTTGGGAGGATAAGCTTTACACGTGCCCCACAAAG
Dragon3 AAGCGAACTGAAGCGGTTGGGAGGATAAGCTTTACACGTGCCCCACAAAG

Import Nexus Data

Now that we’re familiar with the file, let’s import it using the read.nexus.data() function from the ape package:

library(ape)
DragonNexus<-read.nexus.data("Data/DragonMatrix.nex")
head(DragonNexus)

## $`0.1FishXXX`
##  [1] "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0"
## [20] "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0"
## [39] "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0"
## [58] "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0" "0"
## [77] "0" "0"
## 
## $`0.2SnakeXX`
##  [1] "1" "1" "1" "1" "0" "0" "0" "0" "1" "1" "1" "0" "0" "1" "1" "1" "0" "1" "1"
## [20] "0" "0" "0" "0" "0" "0" "0" "1" "1" "0" "0" "0" "0" "0" "0" "1" "0" "0" "0"
## [39] "1" "0" "0" "0" "1" "0" "0" "0" "1" "0" "0" "0" "0" "0" "0" "1" "1" "1" "1"
## [58] "1" "1" "0" "0" "0" "0" "0" "0" "0" "0" "1" "0" "0" "0" "0" "0" "0" "0" "0"
## [77] "0" "0"
## 
## $`0.3MammalX`
##  [1] "1" "0" "0" "0" "0" "0" "0" "0" "0" "0" "1" "1" "1" "0" "0" "0" "0" "0" "0"
## [20] "0" "0" "0" "1" "0" "0" "0" "1" "1" "1" "0" "0" "0" "0" "0" "1" "1" "0" "0"
## [39] "1" "1" "0" "0" "1" "1" "0" "0" "0" "0" "0" "0" "0" "0" "0" "1" "0" "0" "0"
## [58] "0" "0" "0" "1" "1" "0" "0" "0" "0" "0" "1" "0" "0" "0" "0" "0" "1" "0" "0"
## [77] "0" "0"
## 
## $`1GermanXXX`
##  [1] "0" "1" "0" "0" "1" "1" "1" "0" "0" "1" "1" "1" "1" "0" "0" "0" "0" "1" "1"
## [20] "0" "0" "0" "0" "1" "0" "1" "1" "1" "1" "1" "0" "0" "0" "0" "1" "1" "0" "1"
## [39] "0" "0" "0" "0" "1" "1" "0" "1" "1" "0" "0" "0" "0" "0" "0" "1" "1" "1" "0"
## [58] "0" "0" "0" "1" "0" "0" "0" "1" "0" "0" "1" "0" "?" "?" "?" "?" "0" "0" "1"
## [77] "1" "1"
## 
## $`2FrenchXXX`
##  [1] "0" "1" "0" "0" "1" "1" "0" "1" "0" "1" "1" "0" "0" "1" "1" "1" "0" "1" "0"
## [20] "0" "0" "0" "0" "0" "0" "1" "1" "1" "1" "1" "0" "0" "0" "0" "1" "1" "0" "0"
## [39] "1" "0" "0" "0" "1" "1" "0" "0" "0" "0" "0" "0" "0" "0" "0" "1" "1" "0" "0"
## [58] "0" "0" "1" "0" "0" "1" "0" "0" "0" "1" "1" "0" "0" "0" "1" "1" "0" "0" "1"
## [77] "1" "0"
## 
## $`3FrenchXXX`
##  [1] "0" "1" "0" "0" "1" "1" "1" "0" "0" "0" "1" "1" "0" "0" "0" "0" "0" "0" "?"
## [20] "?" "?" "1" "0" "1" "0" "0" "1" "1" "1" "1" "0" "0" "0" "0" "1" "1" "0" "1"
## [39] "1" "1" "0" "1" "1" "1" "0" "1" "0" "0" "0" "0" "0" "0" "0" "1" "1" "1" "0"
## [58] "0" "0" "0" "1" "0" "0" "0" "1" "0" "0" "1" "0" "0" "0" "1" "1" "0" "0" "1"
## [77] "1" "1"

names(DragonNexus)

##  [1] "0.1FishXXX" "0.2SnakeXX" "0.3MammalX" "1GermanXXX" "2FrenchXXX"
##  [6] "3FrenchXXX" "4DutchXXXX" "5EnglishXX" "6AmericanX" "7FrenchXXX"
## [11] "8EnglishXX" "9FrenchXXX" "10FrenchXX" "11SpanishX" "12Japanese"
## [16] "13Japanese" "14Japanese" "15Japanese" "16Japanese" "17Japanese"
## [21] "18Japanese" "19Japanese" "20Japanese" "21Japanese" "22Japanese"
## [26] "23Japanese" "24Japanese" "25Japanese" "26Japanese" "27Japanese"
## [31] "28Japanese" "29Japanese" "30ItalianX" "31ItalianX" "32ItalianX"
## [36] "33XXXXXXXX" "34GermanXX" "35EnglishX" "36GermanXX" "37DutchXXX"
## [41] "38SpanishX" "39ItalianX" "40ItalianX" "41EnglishX" "42ItalianX"
## [46] "43SpanishX" "44ItalianX" "45ItalianX" "46EnglishX" "47ItalianX"
## [51] "48DutchXXX" "49IndianXX" "50Japanese" "51Japanese" "52Japanese"
## [56] "53Japanese" "54IranianX" "55IranianX" "56IranianX" "57IranianX"
## [61] "58TurkishX" "59IranianX" "60IranianX" "61TurkishX" "62TurkishX"
## [66] "63UkraineX" "64UkraineX" "65RussiaXX" "66UkraineX" "67RussiaXX"
## [71] "68GreeceXX" "69ItalianX" "70American" "71BritishX" "72BritishX"
## [76] "73BritishX" "74BritishX"

Compare the header of the nexus R object to the layout of the text-based nexus file. What is different? How does R treat the data?

Distance Matrix

Since we aren’t using DNA, we can’t use the dist.dna() function from ape. Instead, we use the more basic dist() function, which calculates the similarity/dissimilarity matrix based on our binary traits:

DragonDistMat<-dist(DragonNexus,method='binary')

## Error in dist(DragonNexus, method = "binary"): 'list' object cannot be coerced to type 'double'

Why do we get an error?

We get an error because the dist() function doesn’t like the fact that our DragonNexus object is a list. Looking at the ?dist help file tells us what kind of input the function is looking for (look at the description of the x object)

How can we fix this problem?

We can convert a list object to a data.frame object fairly easily, but there is a trick: we need to unlist the list object to make it a vector, before we can convert it to a matrix.

DragonNexusDF<-data.frame(matrix(unlist(DragonNexus), ncol=78,byrow=T))
row.names(DragonNexusDF)<-names(DragonNexus)
head(DragonNexusDF)

##            X1 X2 X3 X4 X5 X6 X7 X8 X9 X10 X11 X12 X13 X14 X15 X16 X17 X18 X19
## 0.1FishXXX  0  0  0  0  0  0  0  0  0   0   0   0   0   0   0   0   0   0   0
## 0.2SnakeXX  1  1  1  1  0  0  0  0  1   1   1   0   0   1   1   1   0   1   1
## 0.3MammalX  1  0  0  0  0  0  0  0  0   0   1   1   1   0   0   0   0   0   0
## 1GermanXXX  0  1  0  0  1  1  1  0  0   1   1   1   1   0   0   0   0   1   1
## 2FrenchXXX  0  1  0  0  1  1  0  1  0   1   1   0   0   1   1   1   0   1   0
## 3FrenchXXX  0  1  0  0  1  1  1  0  0   0   1   1   0   0   0   0   0   0   ?
##            X20 X21 X22 X23 X24 X25 X26 X27 X28 X29 X30 X31 X32 X33 X34 X35 X36
## 0.1FishXXX   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
## 0.2SnakeXX   0   0   0   0   0   0   0   1   1   0   0   0   0   0   0   1   0
## 0.3MammalX   0   0   0   1   0   0   0   1   1   1   0   0   0   0   0   1   1
## 1GermanXXX   0   0   0   0   1   0   1   1   1   1   1   0   0   0   0   1   1
## 2FrenchXXX   0   0   0   0   0   0   1   1   1   1   1   0   0   0   0   1   1
## 3FrenchXXX   ?   ?   1   0   1   0   0   1   1   1   1   0   0   0   0   1   1
##            X37 X38 X39 X40 X41 X42 X43 X44 X45 X46 X47 X48 X49 X50 X51 X52 X53
## 0.1FishXXX   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
## 0.2SnakeXX   0   0   1   0   0   0   1   0   0   0   1   0   0   0   0   0   0
## 0.3MammalX   0   0   1   1   0   0   1   1   0   0   0   0   0   0   0   0   0
## 1GermanXXX   0   1   0   0   0   0   1   1   0   1   1   0   0   0   0   0   0
## 2FrenchXXX   0   0   1   0   0   0   1   1   0   0   0   0   0   0   0   0   0
## 3FrenchXXX   0   1   1   1   0   1   1   1   0   1   0   0   0   0   0   0   0
##            X54 X55 X56 X57 X58 X59 X60 X61 X62 X63 X64 X65 X66 X67 X68 X69 X70
## 0.1FishXXX   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
## 0.2SnakeXX   1   1   1   1   1   1   0   0   0   0   0   0   0   0   1   0   0
## 0.3MammalX   1   0   0   0   0   0   0   1   1   0   0   0   0   0   1   0   0
## 1GermanXXX   1   1   1   0   0   0   0   1   0   0   0   1   0   0   1   0   ?
## 2FrenchXXX   1   1   0   0   0   0   1   0   0   1   0   0   0   1   1   0   0
## 3FrenchXXX   1   1   1   0   0   0   0   1   0   0   0   1   0   0   1   0   0
##            X71 X72 X73 X74 X75 X76 X77 X78
## 0.1FishXXX   0   0   0   0   0   0   0   0
## 0.2SnakeXX   0   0   0   0   0   0   0   0
## 0.3MammalX   0   0   0   1   0   0   0   0
## 1GermanXXX   ?   ?   ?   0   0   1   1   1
## 2FrenchXXX   0   1   1   0   0   1   1   0
## 3FrenchXXX   0   1   1   0   0   1   1   1

DragonDist<-dist(DragonNexusDF,method='binary')

## Warning in dist(DragonNexusDF, method = "binary"): NAs introduced by coercion

DragonDistMat<-as.matrix(DragonDist)

Visualize

To visualize the matrix in ggplot, we need to rearrange the data from an \(n \times n\) matrix to a \(n^2 \times 3\) matrix (i.e. a linear matrix). This is easiliy done with the melt function from the reshape2 library.

library(reshape2)
PDat<-melt(DragonDistMat)

Let’s look at the difference in dimension (structural layout) of the two data objects.

dim(DragonDistMat)

## [1] 77 77

head(DragonDistMat)

##            0.1FishXXX 0.2SnakeXX 0.3MammalX 1GermanXXX 2FrenchXXX 3FrenchXXX
## 0.1FishXXX          0  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX          1  0.0000000  0.7428571  0.6744186  0.6250000  0.7555556
## 0.3MammalX          1  0.7428571  0.0000000  0.6578947  0.7105263  0.6216216
## 1GermanXXX          1  0.6744186  0.6578947  0.0000000  0.5000000  0.2571429
## 2FrenchXXX          1  0.6250000  0.7105263  0.5000000  0.0000000  0.5238095
## 3FrenchXXX          1  0.7555556  0.6216216  0.2571429  0.5238095  0.0000000
##            4DutchXXXX 5EnglishXX 6AmericanX 7FrenchXXX 8EnglishXX 9FrenchXXX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.7435897  0.6829268  0.3870968  0.6888889  0.7948718  0.7368421
## 0.3MammalX  0.7142857  0.7027027  0.7777778  0.7073171  0.9210526  0.6250000
## 1GermanXXX  0.5750000  0.3611111  0.6829268  0.4761905  0.7000000  0.5789474
## 2FrenchXXX  0.3750000  0.4736842  0.6904762  0.4390244  0.6666667  0.4857143
## 3FrenchXXX  0.5135135  0.3055556  0.6666667  0.4883721  0.6829268  0.4166667
##            10FrenchXX 11SpanishX 12Japanese 13Japanese 14Japanese 15Japanese
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6052632  0.7272727  0.6756757  0.6538462  0.4871795  0.4571429
## 0.3MammalX  0.6756757  0.6315789  0.5937500  0.5238095  0.7073171  0.7368421
## 1GermanXXX  0.4871795  0.3243243  0.6097561  0.5000000  0.5681818  0.6279070
## 2FrenchXXX  0.5128205  0.3947368  0.5945946  0.3809524  0.5909091  0.6428571
## 3FrenchXXX  0.4102564  0.2222222  0.6428571  0.4166667  0.5957447  0.6739130
##            16Japanese 17Japanese 18Japanese 19Japanese 20Japanese 21Japanese
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6590909  0.6904762  0.5925926  0.4722222  0.6470588  0.6585366
## 0.3MammalX  0.6052632  0.6923077  0.6400000  0.7435897  0.6000000  0.7105263
## 1GermanXXX  0.5333333  0.5348837  0.6363636  0.6046512  0.5405405  0.5909091
## 2FrenchXXX  0.6222222  0.6190476  0.5483871  0.6190476  0.6388889  0.6511628
## 3FrenchXXX  0.5454545  0.5909091  0.6060606  0.6222222  0.6153846  0.6590909
##            22Japanese 23Japanese 24Japanese 25Japanese 26Japanese 27Japanese
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.7000000  0.7073171  0.6842105  0.6428571  0.6315789  0.6666667
## 0.3MammalX  0.7027027  0.6756757  0.6857143  0.6750000  0.6756757  0.6842105
## 1GermanXXX  0.5365854  0.5609756  0.5500000  0.5869565  0.5581395  0.5250000
## 2FrenchXXX  0.6250000  0.5853659  0.6410256  0.5714286  0.5500000  0.6136364
## 3FrenchXXX  0.5952381  0.6000000  0.6097561  0.6000000  0.5365854  0.6000000
##            28Japanese 29Japanese 30ItalianX 31ItalianX 32ItalianX 33XXXXXXXX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.7021277  0.6829268  0.6666667  0.7073171  0.6000000  0.7500000
## 0.3MammalX  0.7500000  0.7027027  0.6666667  0.5757576  0.6486486  0.7894737
## 1GermanXXX  0.5555556  0.5952381  0.4871795  0.3611111  0.3947368  0.4444444
## 2FrenchXXX  0.5957447  0.5609756  0.4166667  0.5128205  0.4750000  0.4285714
## 3FrenchXXX  0.6041667  0.6136364  0.4615385  0.2857143  0.3243243  0.4054054
##            34GermanXX 35EnglishX 36GermanXX 37DutchXXX 38SpanishX 39ItalianX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6666667  0.7446809  0.7826087  0.6739130  0.6666667  0.6829268
## 0.3MammalX  0.7142857  0.8222222  0.6842105  0.7209302  0.6829268  0.7027027
## 1GermanXXX  0.4146341  0.5116279  0.4750000  0.3902439  0.4523810  0.6046512
## 2FrenchXXX  0.4523810  0.4761905  0.4358974  0.5000000  0.4146341  0.4473684
## 3FrenchXXX  0.3902439  0.5555556  0.3684211  0.2564103  0.3500000  0.5714286
##            40ItalianX 41EnglishX 42ItalianX 43SpanishX 44ItalianX 45ItalianX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6976744  0.7567568  0.7209302  0.7045455  0.6857143  0.6363636
## 0.3MammalX  0.6486486  0.6666667  0.6756757  0.6578947  0.6129032  0.6829268
## 1GermanXXX  0.3947368  0.5000000  0.3333333  0.4102564  0.4411765  0.3750000
## 2FrenchXXX  0.4750000  0.4545455  0.4210526  0.4500000  0.4411765  0.4883721
## 3FrenchXXX  0.3243243  0.4411765  0.2777778  0.3846154  0.4117647  0.2631579
##            46EnglishX 47ItalianX 48DutchXXX 49IndianXX 50Japanese 51Japanese
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.7179487  0.7333333  0.7272727  0.5945946  0.6578947  0.6595745
## 0.3MammalX  0.6060606  0.6923077  0.6842105  0.6060606  0.6571429  0.7446809
## 1GermanXXX  0.5263158  0.4500000  0.5238095  0.5952381  0.5384615  0.4666667
## 2FrenchXXX  0.4722222  0.4500000  0.3947368  0.5526316  0.5263158  0.3414634
## 3FrenchXXX  0.3888889  0.3421053  0.4250000  0.6000000  0.5714286  0.4222222
##            52Japanese 53Japanese 54IranianX 55IranianX 56IranianX 57IranianX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6176471  0.5862069  0.7073171  0.5789474  0.6875000  0.6511628
## 0.3MammalX  0.5666667  0.6206897  0.6756757  0.6111111  0.7333333  0.6315789
## 1GermanXXX  0.4722222  0.5714286  0.4736842  0.6046512  0.5531915  0.6000000
## 2FrenchXXX  0.5555556  0.4193548  0.4473684  0.5250000  0.4186047  0.5000000
## 3FrenchXXX  0.6000000  0.4857143  0.3421053  0.5238095  0.4666667  0.5777778
##            58TurkishX 59IranianX 60IranianX 61TurkishX 62TurkishX 63UkraineX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.4750000  0.6923077  0.6585366  0.6410256  0.5750000  0.6744186
## 0.3MammalX  0.6410256  0.7500000  0.6388889  0.6571429  0.6578947  0.5833333
## 1GermanXXX  0.6170213  0.6976744  0.6666667  0.6363636  0.5454545  0.5238095
## 2FrenchXXX  0.6086957  0.6341463  0.6363636  0.6000000  0.5365854  0.6222222
## 3FrenchXXX  0.5744681  0.6744186  0.5476190  0.6097561  0.5476190  0.5681818
##            64UkraineX 65RussiaXX 66UkraineX 67RussiaXX 68GreeceXX 69ItalianX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6388889  0.6000000  0.7000000  0.5862069  0.6976744  0.7777778
## 0.3MammalX  0.5483871  0.6111111  0.6666667  0.5384615  0.6842105  0.7750000
## 1GermanXXX  0.6250000  0.5813953  0.4324324  0.5588235  0.5813953  0.6046512
## 2FrenchXXX  0.5789474  0.4358974  0.5384615  0.4516129  0.2571429  0.4615385
## 3FrenchXXX  0.5128205  0.5000000  0.3333333  0.5294118  0.4634146  0.5238095
##            70American 71BritishX 72BritishX 73BritishX 74BritishX
## 0.1FishXXX  1.0000000  1.0000000  1.0000000  1.0000000  1.0000000
## 0.2SnakeXX  0.6388889  0.7441860  0.6666667  0.6744186  0.7021277
## 0.3MammalX  0.5937500  0.7692308  0.8000000  0.7250000  0.7500000
## 1GermanXXX  0.4722222  0.4358974  0.5744681  0.5116279  0.4888889
## 2FrenchXXX  0.4571429  0.5000000  0.6000000  0.5000000  0.4761905
## 3FrenchXXX  0.4736842  0.4871795  0.5116279  0.4750000  0.4523810

dim(PDat)

## [1] 5929    3

head(PDat)

##         Var1       Var2 value
## 1 0.1FishXXX 0.1FishXXX     0
## 2 0.2SnakeXX 0.1FishXXX     1
## 3 0.3MammalX 0.1FishXXX     1
## 4 1GermanXXX 0.1FishXXX     1
## 5 2FrenchXXX 0.1FishXXX     1
## 6 3FrenchXXX 0.1FishXXX     1

Now let’s plot. Note: Here is a good resource from the developers of ggtree. This GitHub link includes details on how to make a visually appealing tree.

library(ggplot2)
ggplot(data = PDat, aes(x=Var1, y=Var2, fill=value)) + 
  geom_tile()+scale_fill_gradientn(colours=c("white","blue","green","red")) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5))

Looks like we have a nice range of values to try to cluster by distance.

Tree Building

Now that we have a distance matrix, let’s try building a phylogeny using the neighbour-joining (NJ) method.

DragonTree<-nj(DragonDist)

Now let’s draw the tree using the ggtree library to compare the two:

devtools::install_github("hadley/devtools")

## rlang       (1.0.1 -> 1.0.2  ) [CRAN]
## glue        (1.6.1 -> 1.6.2  ) [CRAN]
## withr       (2.4.3 -> 2.5.0  ) [CRAN]
## desc        (1.4.0 -> 1.4.1  ) [CRAN]
## xfun        (0.29  -> 0.30   ) [CRAN]
## sass        (NA    -> 0.4.0  ) [CRAN]
## bslib       (NA    -> 0.3.1  ) [CRAN]
## textshaping (NA    -> 0.3.6  ) [CRAN]
## openssl     (1.4.6 -> 2.0.0  ) [CRAN]
## promises    (NA    -> 1.2.0.1) [CRAN]
## later       (NA    -> 1.3.0  ) [CRAN]
## sourcetools (NA    -> 0.1.7  ) [CRAN]
## fontawesome (NA    -> 0.2.2  ) [CRAN]
## httpuv      (NA    -> 1.6.5  ) [CRAN]
## rmarkdown   (2.11  -> 2.12   ) [CRAN]
## ragg        (NA    -> 1.2.2  ) [CRAN]
## downlit     (NA    -> 0.4.0  ) [CRAN]
## shiny       (NA    -> 1.7.1  ) [CRAN]
## crosstalk   (NA    -> 1.2.0  ) [CRAN]
## urlchecker  (NA    -> 1.0.1  ) [CRAN]
## profvis     (NA    -> 0.3.7  ) [CRAN]
## pkgdown     (NA    -> 2.0.2  ) [CRAN]
## miniUI      (NA    -> 0.1.1.1) [CRAN]
## DT          (NA    -> 0.21   ) [CRAN]
## 
##   There is a binary version available but the source version is later:
##         binary source needs_compilation
## pkgdown  1.6.1  2.0.2             FALSE
## 
## package 'rlang' successfully unpacked and MD5 sums checked
## package 'glue' successfully unpacked and MD5 sums checked
## package 'withr' successfully unpacked and MD5 sums checked
## package 'desc' successfully unpacked and MD5 sums checked
## package 'xfun' successfully unpacked and MD5 sums checked
## package 'sass' successfully unpacked and MD5 sums checked
## package 'bslib' successfully unpacked and MD5 sums checked
## package 'textshaping' successfully unpacked and MD5 sums checked
## package 'openssl' successfully unpacked and MD5 sums checked
## package 'promises' successfully unpacked and MD5 sums checked
## package 'later' successfully unpacked and MD5 sums checked
## package 'sourcetools' successfully unpacked and MD5 sums checked
## package 'fontawesome' successfully unpacked and MD5 sums checked
## package 'httpuv' successfully unpacked and MD5 sums checked
## package 'rmarkdown' successfully unpacked and MD5 sums checked
## package 'ragg' successfully unpacked and MD5 sums checked
## package 'downlit' successfully unpacked and MD5 sums checked
## package 'shiny' successfully unpacked and MD5 sums checked
## package 'crosstalk' successfully unpacked and MD5 sums checked
## package 'urlchecker' successfully unpacked and MD5 sums checked
## package 'profvis' successfully unpacked and MD5 sums checked
## package 'miniUI' successfully unpacked and MD5 sums checked
## package 'DT' successfully unpacked and MD5 sums checked
## 
## The downloaded binary packages are in
##  C:\Users\rob_c\AppData\Local\Temp\RtmpK6cGb4\downloaded_packages
## * checking for file 'C:\Users\rob_c\AppData\Local\Temp\RtmpK6cGb4\remotes25884f235da9\r-lib-devtools-575ae4e/DESCRIPTION' ... OK
## * preparing 'devtools':
## * checking DESCRIPTION meta-information ... OK
## * checking for LF line-endings in source and make files and shell scripts
## * checking for empty or unneeded directories
## * building 'devtools_2.4.3.9000.tar.gz'
## 

library(ggtree)
ggtree(DragonTree,layout="circular")

ggtree(DragonTree,layout="rectangular")

Woah, what’s going on here?

This tree has some problems. The branches are very long relative to the bifurcations among groups. It is almost as if all the characters are all mixed up. This is what we might expect if dragons were created from active imaginations and didn’t really evolve from each other.

Another reason is that we are treating all traits the same. For example, we treat snout length the same as limb number. However, we might argue that limb length evolves more slowly than snout length, that dragons with the same number of limbs are more likely related than dragons with similar snout lengths. We can do this by weighting our traits. Weights are just numbers that we multiply (or sometimes add) to our data so that traits with higher weights have a stronger influence on our clustering algorithm.

Weights

The Weights.csv data table has a set of weights that were used for the Dragon Phylogeny t-shirt design. Let’s take a look:

WeightsDat<-read.csv("Data/Weights.csv")

Code	Origin	Weight	Rationale
1	Appendages	ZZZZ	Highly Highly conserved (all tertrapods came from the same fish; loss of limbs more common (e.g. lizards to snakes)
2	Mass	1111	Weak: e.g. squirrel vs. Kangaroo
3	Body Type	CC	Body type somewhat conserved (e.g. snakes vs. lizards)
4	ClawType	5555	Somewhat Strong: e.g. mammals vs. eagle
5	Dorsal Ridges	111111	Weak: e.g. dinosaurs
6	Ear Morphology	111	Weak: e.g. deer mouse vs. vole
7	Eye Morphology	111	Weak: e.g. nocturnal vs. diurnal and eye size
8	Eye Position	1	Weak: e.g. predators vs. prey
9	HornType	111	Weak: e.g. gazelle vs. deer
10	NosePos	7	Moderately conserved e.g. amphibians vs. fish
11	Nasal Morphology	1	Weak: e.g. dog breeds
12	Skin-dorsal	999999	Somewhat conserved: e.g. reptiles vs. mammals
13	Skin-head	333333	Somewhat conserved: e.g. reptiles vs. mammals
14	Skin-ventral	333333	Somewhat conserved: e.g. reptiles vs. mammals
15	SnoutType	3333	Moderate: e.g. mouse vs. rhinocerous
16	TailType	11	Weak: e.g. dinosaurs
17	Teeth	3333	Moderately conserved e.g. herbivore vs. carnivore mammals
18	OppToes	1	Weak: e.g. chimps vs. humans
19	Toe Number	333333	Moderate: e.g. horse vs. lion
20	Tongue Length	1	Weak: e.g. anteater vs. primate
21	Tongue Morphology Type	44	Moderate: e.g. snake vs. lizard
22	Ventral Plates	7	Somewhat conserved: e.g. snakes vs. mammals
23	Wiskers	11	Weak: e.g. catfish vs. bass
24	Wing Structure	JJ	Somewhat highly conserved; single origin in bats and birds
25	WingType	AAA	Strong: e.g. bat vs. bird

Weights are scaled on a scale from 1 through 9 and then A (A=10) through Z (Z=35). The number of each digit corresponds to the number of binary values for that trait, and the traits are ordered in the same way they were encoded. There are programs we could use to calculate distance using these as weights (e.g. BEAST2). However, we’ll do it manually to see how it works. All we need to do is multiply each binary value by it’s weight. So the first step is to import the weights and extract the weights column. However, that’s going to take a few steps:

1. Create a single vector of weights
2. Convert each letter to its corresponding weight value (e.g. A=10, B=11, etc.)
3. Multiply the weight value by the trait vector for each dragon
4. Re-calculate our distance matrix
5. Plot the tree

1. Create a single vector of weights.

Easy:

Weights<-paste0(WeightsDat$Weight,collapse="")
Weights<-strsplit(Weights,split="")[[1]]

2. Convert each letter to a value.

We could encode very single letter individually, or we can use a custom function with the built-in LETTERS object:

LETTERS # See what LETTERS is (see also letters)

##  [1] "A" "B" "C" "D" "E" "F" "G" "H" "I" "J" "K" "L" "M" "N" "O" "P" "Q" "R" "S"
## [20] "T" "U" "V" "W" "X" "Y" "Z"

which(LETTERS=="G")

## [1] 7

WeightsNum<-rep(NA,length(Weights))
for(i in 1:length(WeightsNum)){
  if(Weights[i] %in% LETTERS){
    WeightsNum[i]<-which(LETTERS==Weights[i])+9
  } else {
    WeightsNum[i]<-Weights[i]
  }
}
WeightsNum<-as.numeric(WeightsNum)

Now we have a vector of weights, which should have the same length as the number of 1s and 0s (i.e. 78 ‘characters’):

length(WeightsNum)

## [1] 78

3. Multiply the weight value by the trait vector for each dragon.

This is complicated by the fact that our data include missing data coded as ?, so all of our characters are stored as strings. Since \(0*x=0\) and \(?*x\) is undefinded, we really just need to multiply the 1s, which is equivalent to replacing the 1s with their corresponding weight values. To do this, we also have to slice our list object using the double brackets [[]].

WtDragonNexus<-DragonNexus # Make a new weighted data frame object
for (i in 1:length(DragonNexus)){
  RepWeight<-DragonNexus[[i]]==1
  WtDragonNexus[[i]][RepWeight]<-WeightsNum[RepWeight]
  RepWeight<-NA
}

4. Re-calculate our distance matrix

We just need to modify our objects

WtDragonNexusDF<-data.frame(matrix(unlist(WtDragonNexus),ncol=78,byrow=T))
row.names(WtDragonNexusDF)<-names(WtDragonNexus)
WtDragonDist<-dist(WtDragonNexusDF,method='euclidean')

## Warning in dist(WtDragonNexusDF, method = "euclidean"): NAs introduced by
## coercion

WtDragonDistMat<-as.matrix(WtDragonDist)

Note the change in method from binary to euclidean… why?

Compare the new distance matrix to the older one above. Note the much stronger structure:

WtPDat<-melt(WtDragonDistMat)
ggplot(data = WtPDat, aes(x=Var1, y=Var2, fill=value)) + 
  geom_tile()+scale_fill_gradientn(colours=c("white","blue","green","red")) +
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5))

5. Plot the tree

Let’s compare the minimal evolution (ME) and Neighbour Joining (NJ) methods

WtDragonTree<-fastme.bal(WtDragonDist)
WtDragonTreeNJ<-nj(WtDragonDist)
ggtree(WtDragonTree,layout="circular")

ggtree(WtDragonTreeNJ,layout="circular")

Tree Formatting

Let’s try to make it look a bit better. To do this, we need to understand the data format

str(WtDragonTree)

## List of 4
##  $ edge       : int [1:151, 1:2] 78 78 79 80 81 81 82 83 83 82 ...
##  $ edge.length: num [1:151] 25.6 5.88 9.8 1.63 12.13 ...
##  $ tip.label  : chr [1:77] "0.1FishXXX" "8EnglishXX" "32ItalianX" "73BritishX" ...
##  $ Nnode      : int 75
##  - attr(*, "class")= chr "phylo"
##  - attr(*, "order")= chr "cladewise"

Note that this is a list object with phylo class and cladewise order. We can also see three ‘slices’ (denoted with $). The tip.label slice contains the specimen labels. The edge slice contains all of the line segments,

Q: Why are there 151 edges if we only have 77 dragons?

The edges include lines that connect clades.

Colour by OTU

The tip labels contain information about the origin of each dragon image (‘Fish’, ‘Reptile’ and ‘Mammal’ were added later as outgroups).

head(WtDragonTree$tip.label)

## [1] "0.1FishXXX" "8EnglishXX" "32ItalianX" "73BritishX" "37DutchXXX"
## [6] "45ItalianX"

We can use this to colour-code our tree and see if dragons from the same regions cluster together. We can use regular expressions to parse out a vector:

Country<-gsub("[0-9\\.]+([^X]+)X*","\\1",WtDragonTree$tip.label) # Remove leading numbers

Next we have to group tip.labels by their corresponding country. There is a nice function in R called split that makes this easy to do:

CountryGroups<-split(WtDragonTree$tip.label, Country)

Now we use the groupOTU function to apply the grouping information for plotting:

WtDTcol<-groupOTU(WtDragonTree,CountryGroups)
str(WtDTcol)

## List of 4
##  $ edge       : int [1:151, 1:2] 78 78 79 80 81 81 82 83 83 82 ...
##  $ edge.length: num [1:151] 25.6 5.88 9.8 1.63 12.13 ...
##  $ tip.label  : chr [1:77] "0.1FishXXX" "8EnglishXX" "32ItalianX" "73BritishX" ...
##  $ Nnode      : int 75
##  - attr(*, "class")= chr "phylo"
##  - attr(*, "order")= chr "cladewise"
##  - attr(*, "group")= Factor w/ 19 levels "3","American",..: 6 5 12 3 4 12 8 17 3 11 ...

Notice how there is a new group attribute, which is a factor containing our country groups.

ggtree(WtDTcol,layout="circular",aes(colour=group))+geom_tiplab(size=2,aes(angle=angle))

What might we infer from this figure?

Colour by clade

As an alternative to colouring by region, we might want to point out a few clades (i.e. groups of dragons that cluster together). For example, it looks like the outer node

WtDTclade<-groupClade(WtDragonTree,.node=c(142,128,103,90,80))
ggtree(WtDTclade,layout="circular",aes(colour=group)) + 
  geom_cladelabel(node=142,label="Serpentidae",hjust=0.5,offset.text=4,fontsize=3,angle=-45) +
  geom_cladelabel(node=128,label="Wyvernidae",hjust=0.5,offset.text=4,fontsize=3,angle=15) +
  geom_cladelabel(node=103,label="Orientalia",hjust=0.5,offset.text=4,fontsize=3,angle=40) +
  geom_cladelabel(node=90,label="Dracopteronidae",hjust=0.5,offset.text=4,fontsize=3,angle=-55) +
  geom_cladelabel(node=80,label="Dracoverisidae",hjust=0.5,offset.text=6,fontsize=3,angle=55) +
  xlim(NA,60)

NOTE: To find these nodes, we can use + geom_nodelab to plot the node number on top of each node.

Advanced Techniques:

(OPTIONAL) You can Try overlaying your phylogeny on a geographical map: https://www.molecularecologist.com/2014/11/geophylogeny-plots-in-r-for-dummies/

To do this, you would need to find latitude/longitude coordinates. An easy way to do this is to find a location in Google Maps. When you right-click you will see the longitude and latitude coordinates.