A BIOGEOGRAPHICAL CASE STUDY OF DIDERMA HEMISPHAERICUM (MYXOMYCOTA) predicting local habitat suitability in changing climate scenarios and assessing genetic diversity across southeast asian populations

THAI NGUYEN UNIVERSITY UNIVERSITY OF AGRICULTURE AND FORESTRY KING JOSHUA ALMADRONES REYES A BIOGEOGRAPHICAL CASE STUDY OF DIDERMA HEMISPHAERICUM MYXOMYCOTA: PREDICTING LOCAL HABITAT

THAI NGUYEN UNIVERSITY

UNIVERSITY OF AGRICULTURE AND FORESTRY

KING JOSHUA ALMADRONES REYES

A BIOGEOGRAPHICAL CASE STUDY OF

DIDERMA HEMISPHAERICUM (MYXOMYCOTA):

PREDICTING LOCAL HABITAT SUITABILITY IN CHANGING CLIMATE SCENARIOS AND ASSESSING GENETIC DIVERSITY ACROSS SOUTHEAST

ASIAN POPULATIONS

Study Mode: Full-time

Faculty: Advanced Education Program Office

BACHELOR THESIS

Thai Nguyen, 15/09/2018

Thai Nguyen University of Agriculture and Forestry

Degree Program Bachelor of Science in Environmental Science and

Management

DIDERMA HEMISPHAERICUM (MYXOMYCOTA):

Predicting local habitat suitability in changing climate scenarios

and assessing genetic diversity across Southeast Asian populations

Supervisor(s) Dr.rer.nat Nikki Heherson A Dagamac & Mr Do Tuan Tung Abstract: Biogeographic and molecular studies are currently the trends for biodiversity assessments to address issues related to conservation and biological resource management Since these areas are underexplored for economically important terrestrial protist like the myxomycetes, two research

frameworks using the cosmopolitan Diderma hemisphaericum as the

biological model were designed for this study: (1) predicting suitable niches for the species under different climate change scenarios and (2) assessing genetic diversity of the species population between Philippines and Vietnam using partial 18s rDNA gene marker The first framework showed: (i) temperature seasonality and isothermality to be the most important

bioclimatic predictors and (ii) widespread increase of suitable habitats for D

hemisphaericum in A2 and B1 scenarios This is the first report of species

distribution modeling for the Southeast Asian myxomycetes which can be used for finding priority areas for conservation in the Philippines The second framework revealed: (i) that gene flow exist yielding a total of 27 ribotypes coming from 4 suspected putative biospecies and (ii) geographical barriers are

not the most significant driver in the speciation events for the D

hemisphaericum populations supporting now the “Everything is everwhere”

model of biogeographic distribution This finding provides groundbase information for myxomycetes biodiversity at the genetic level in Southeast Asian region

Keywords: Biogeography –climate change –environmental niche

–fungus-like protists –molecular diversity Number of pages: 102

Date of

Submission: September 15, 2018

ACKNOWLEDGEMENT

The King would like to give its deepest gratitude to the people who made this one hell

of a scientific work:

 To Dr rer nat Nikki Heherson Aldea Dagamac (University of Greifswald) and

my Vietnamese adviser Mr Do Tuan Tung (TUAF) for their supervision

 To Prof Dr Martin Schnittler and the International Office in Greifswald, thank you very much for the financial assistance during my research internship in Greifswald, Germany

 To the AEP Office, especially Chi Hong, thank you very much for all the assistance you’ve extended from Visa applications and all the questions you

enthusiastically entertains Also, huge thanks to my roommate Tu Quang Tu

 To Tita Dory Zapanta and Tito Joey Zapanta, thank you for all the support you have been extending to me and my family from my high school years until now that whom without, I will be probably in a very different path

 To Tita Marge Parulan, thank you very much for your good heart and support for

me and my family And to all the many people who helped me and my family through thick and thin, thank you very much!

 I would also like to give thanks to the General Botany and Plant Systematics working group of the Institute of Botany and Landscape Ecology for the scientific discussions, barbecue parties and table tennis on a very hot German summer –

Frau Anja Klahr, Dr Manuela Bog, Oleg Shchepin, Jan Woyzichovski, Melanie Zacharias and Paul Lamkowski

 Huge thanks to my Lola Flor, Kuya Etang, Kuya Eric and Kuya Erwin for always checking my condition everytime

 To the Hood - Niecer, JD, Aaron and Ian thank you for not forgetting to send memes during my stay in Germany, I really appreciate yall doing that

 A huge shoutout to Nami, Blindie and Kaneki You’re my favorite pets that kept

me smiling even though I am far away from Home

 To Yani and Enzo – we survived the MyxoGod

 I would also like to give thanks to the people that I’ve met in Greifswald - Kasia,

Oleg, Jan, Mori, Virna, Nadine, Felix 1, Linh, Felix 2, Benjamin, Julia, Jonas, Lukas, Christoph, Paul, Robin, Oriana, Andrea, Laura 1, Laura 2, Prakash -

for the parties, hotpot, picnics, volleyball and board game nights we shared Vielen dank for making me feel at home in that one small medieval-ish town See

ya fellow adventurers and guardians (I have 2 golds)!

 Also, very big thanks to Sir Tom, Ate Jeane and Baby Jillian dela Cruz, for the warm welcome in Berlin

 Next, to my Asian friends, Lorna, Lita, and Fibi, thanks for the Asian company and for the reminder to try “new” things in life

 Moreover, to these people, first, Nikki, thank you for the memorable experience while hanging out in Europe I really enjoyed the sisig you cooked while staying in

your flat (still it’s my recipe) Second, to William, thanks for the fun while

travelling in Prague and continously telling me to go back to my island Also to

Ward, who always have good jokes (well, sometimes), thank you for reminding me

to behave like I’m not Thanks to the three of you for making my Greifswald (and Domburg) experience complete See you all next time and let’s go to a church someday

 Also to my partner, Maria Yssabella Sinfuego You have been my fuel and inspiration Thank you for all the times you checked my level of insanity and your never ending encouragement and support from the start

 To my MENTOR, I couldn’t express the gratitude for all the guidance and patience you gave me during my stay in Germany I have learned so many things, not just in the field of science but also about the many perspectives in life Lastly, thank you for believing and seeing the potential in me when no one else can

This masterpiece is dedicated to my supportive parents:

My mom, Jocelyn Almadrones Reyes, and My dad, Efren Martin Reyes

Your unconditional love all throughout my journey of ups and down made me the person I

am right now I am blessed that I was raised by the best parents in the world who has been my inspiration to every life decisions I need to make and I will make in the future

I hope I can make you both proud on the MAN that I have become!

“You can do anything you set your mind to”

- Eminem

“Jesus loves you more than you will know”

- King 2k18

TABLE OF CONTENTS

List of Figures 8

List of Tables 9

List of Abbreviations 10

Part I Introduction 11

1.1 Research Rationale 11

1.2 Research Questions and Hypotheses 14

1.2.1 Predicting local habitat suitability in changing climate scenarios 14

1.2.2 Assessing gene flow of Diderma hemisphaericum across regional geographical barrier 15

1.3 Research Objectives 17

1.4 Scope and Limitations 18

1.5 Definition of terms 19

Part II Literature Review 22

2.1 Biogeography of myxomycetes 22

2.2 Species Distribution Modelling (SDM) 25

2.3 DNA Barcoding 28

Part III Methodology 32

3.1 Predicting habitat suitability in changing climate snenario 32

3.1.1 Gathering the distribution records of D hemisphaericum in the Philippines 32

3.1.2 Obtaining current and future (A2 and B1) environmental layers 32

3.1.3 Species Distribution Modelling 33

3.2 Assessing genetic diversity across regional populations 35

3.2.1 Obtaining samples of Diderma hemisphaericum 35

3.2.2 DNA extraction, amplification and sequencincing 35

3.2.3 Quality filtering and sequence alignment 37

3.2.4 Phylogenetic tree and ribotype networking 38

3.2.5 Analysis of the genetic diversity 39

Part IV Results and Discussion 40

4.1 Predicting local habitat suitability in changing climate scenarios 40

4.1.1 Results 40

4.1.2 Discussion 44

4.2 Assessing genetic diversity across regional populations 46

4.2.1 Results 46

4.2.2 Discussion 55

PART V Summary, Conclusion and Recommendation 62

References 67

Appendices 83

Figure 1: Fructification of Diderma hemisphaericum viewed under a compound stereo microscope 13

Figure 2: A diagram that shows major steps for constructing a robust correlative species distribution model 34

Figure 3: Species distribution models for Diderma hemisphaericum showing map of the

Philippines and the predictive suitable habitat areas under current conditions and two future changing climate scenarios for the year 2080 43 Figure 4: The rooted consensus tree based on the maximum likelihood algorithm constructed on successfully amplified DNA products from specimens coming from Philippines and Vietnam based on partial 18S rDNA sequences The scale bar indicates the evolutionary distance per site and the black dot indicates the branch Grey triangles indicate the collapsed specimens having the same ribotype assignment 50 Figure 5: The constructed ribotype network constructed using TCS that also showed division into four different clades The circle size indicates the number of sequences represented by the ribotypes The connecting line segments indicate mutational steps between alleles and the small circle in between represent the hypothetical transitional ribotypes 51 Figure 6: Species accumulation curve based on the Chao 1 estimates in (A) Vietnam and (B) Philippines The dark blackline shows the Chao 1 mean and the grey highlight indicates the 5% and 95% limits (C) A smooth species accumulation curve based on

Coleman rarefaction 57

LIST OF TABLES

Table 1: Number of species occurrence record used for the models, AUC- Values, percentage contribution, permutation importance and training gain (with and without respective variable) for the 5 most contributory predictors 42 Table 2: Summary of the polymorphic site analysis comparing the two countries generated by the DNAsp v.5.1: (i) Philippines and (ii) Vietnam 47 Table 3: AMOVA (Analaysis of molecular variance) displays the computed FST index (0.083), with (0) indicating genetic material sharing and (1) as no sharing 47

Table 4: Correlation using mantel test between the genetic and geographic distance of the

samples from Philippines and Vietnam 49

Table 5: List of samples of Diderma hemisphaericum with their corresponding; location,

collector, ribotype number, and corresponding coordinates 52

LIST OF ABBREVIATIONS

AMOVA Analysis of Molecular Variance BLAST Basic local alignment search tool

EF1a Elongation factor 1 alpha gene

GBIF Global Biodiversity Information Facility MAFFT Multiple alignment using fast Forier transform

MEGA Molecular evolutionary genetic analysis PCR Polymerase chain reaction

SDM Species distribution modelling

TAE Tris base, acetic acid, and EDTA

Myxomycetes or the true slime molds, belong to the supergroup of Amoebozoa that are commonly found in many terrestrial ecosystem where there are decaying organic substrates (Stephenson & Kahlert, 2018; Lado et al., 2003) Most of their life stage, they are in the amoebal forms which came from a germinated spores that are dispersed in the

to become a zygote that will lead to the development of this giant mass of multinucleated cell called “the blob” or scientifically known as the plasmodia This plasmodial stage of the myxomycetes is known to represent the animal feature in its whole life stage since this

is the structure that can move across any substrate using cytoplasmic streaming (Kamiya, 1968) Furthermore, the plasmodia served as a limiting factor in the ecosystem because they act as microbial predators in the soil that eat bacterial and fungal species maintaining now the ecological balance in nature (Stephenson & Schnittler, 2017) When the food source in the surrounding of the plasmodia becomes limited and the environmental condition started to become unfavorable for them, these intelligent blobs starts to signal themselves to transform from an animal like feature to becoming a fungal fruiting body These macroscopic structures of the fruiting bodies are what fascinated many myxomyceteologist around the world These fascinations led to researches that focus on fructifications-based ecological studies (Ronikier & Ronikier, 2009; Martin, 1940) and global pattern distributions (Dagamac et al., 2017a, b & c; Rollins & Stephenson, 2011)

In spite of the increasing number of studies about myxomycetes, most of these studies are concentrated in the temperate ecozones and only just recently that the hype in conducting studies in the tropical regions begun, in particular in the less explored Southeast Asian (SEA) Palaeotropical ecozones With the many information that started to fill in the

missing gaps about the global distribution of the myxomycetes in SEA, (Philippines, Dagamac et al., 2014, Thailand, Tran et al., 2006, Vietnam, Novozhilov et al., 2017) it is

deemed necessary to start applying now some of the many modern technologies that are used for conservation such as modeling the environmental niche distribution or detecting

the gene flow of species population across different countries An ideal myxomycete species that can serve as a flagship model to start this baseline information should be at least (1) abundantly occurring in most SEA collections, and (2) clear cut and easily identifiable species that even a layman person could determine From these purposes, a

good candidate is the common tropical myxomycetes but less explored model, Diderma

hemisphaericum (Bull.) Hornem

Figure 1 Fructification of Diderma hemisphaericum viewed under a compound stereo microscope

Diderma hemisphaericum is a “chocolate crinkle” looking myxomycetes species that

was first found in the article of Bulliard Hornemann in 1829 in a book called Flora

Danica This species is reported in the temperate countries such as Europe according to

the Global Biodiversity Information Facility (GBIF) web portal and studies of Chen et al.,

2013 and Ing (1999) However, recent studies in Palaeotropical forests in Southeast Asian

countries (see studies of Redeña-Santos et al., 2017; Tran et al., 2014; Rosing et al., 2011;

Ko Ko et al., 2010), shows that D hemisphaericum can also thrive and effortlessly found

in a tropical and subtropical climate In summary, it can be assumed that D

hemisphaericum can be a subject matter in global scale dispersal Therefore, it is the best

candidate for the two research frameworks in this study

1.2 Research Questions and Hypotheses

Since climate change is happening, biological responses, in particular with microorganism, are necessary to be established The initial step for this is to understand biogeographical distribution of species Given the conjecture that local and regional

distribution of Diderma hemisphaericum still remains as a conundrum in Southeast Asia,

this study was designed to have two research frameworks that employed two different but highly related methodologies that addresses questions in many conservation studies like local range dynamics where a species can productively occur and the influence of regional scale geographical barriers in a clear-cut population of myxomycetes

1.2.1 Predicting local habitat suitability in changing climate scenarios

Background: Modeling the species distribution of an organisms are becoming a recent

trend in many biogeographical studies This now leads to many recent publication in myxomycetes that employed the use of such technique (see studies of Rojas et al., 2015)

for conservation purposes In the case of Diderma hemisphaericum, even though they are

the abundantly occurring myxomycetes in most local myxomycetes studies in the

Philippines, there is relatively little information about the local distribution of this species

in the Philippines where myxomycete studies had been undergoing during the last four decades due to the fact that most occurrence data that are available as of this date are only coming from extensive studies in selected Northern islands of the country

Questions: Hence, to address this predicament, the following research questions were

formulated:

a) What are the other possible areas or suitable habitats in the Philippine archipelago

where D hemisphaericum can be found using the maximum entropy algorithm?

b) How would be the predictive local distribution of D hemisphaericum after subjecting

them to two (A2 and B1) different climate change scenarios?

Hypothesis: Since environmental conditions in other areas in the Philippines are expected

to occur at least the same at a local scale, the species distribution models (both for the current and changing climate scenarios) generated in this study would support the

cosmopolitan distribution hypothesis for D hemisphaericum

1.2.2 Assessing gene flow of Diderma hemisphaericum across regional geographical barriers

Background: In recent biogeographical studies, DNA barcoding is one of the key in

solving the two proposed hypothesis of protist biogeography which is the (i) Everything is Everywhere (EiE) model and (ii) Moderate Endemicity (ME) model The EiE model

suitable for them While the latter ME model, suggests that species have different DNA

“barcode” (biospecies) due to the fact that there are geographical barriers limiting the distribution of the species A particular study of Dagamac et al., 2017a, a myxomycete

species namely Hemitrichia serpula, was used to address the two opposing hypothesis

Questions: Thus, for this study, populations of Diderma hemisphaericum from two

regional regions in Philippines and Vietnam were used to answer the following questions:

a) Do genetic clades of D hemisphaericum populations show gene flow between

Philippines and Vietnam?

b) Does geographic barriers play a role in the distribution of genetic ribotypes at a regional biogeographic scale?

Hypotheses: If the null hypothesis that states that there is no significant difference

between the gene flow of Philippines and Vietnam population of D hemisphaericum is

accepted, then the Everything is Everywhere (EiE) model of protist biogeography is true However, if the alternative hypothesis that is phrased as; clear bifurcation of gene sequences between Philippines and Vietnam samples, then geographical barriers like

oceans and mountain ranges, affect the gene flow among populations of D

hemisphaericum, supporting now the Moderate Endemicity (ME) model

On the other hand, the following aims were constructed to respond to the questions raised

in the second framework of this Bachelor thesis:

◉ to extract, amplify, visualize, purify and sequence DNA from D hemisphaericum

samples from Vietnam and Philippines

◉ to obtain partial sequences of 18S rDNA (SSU) barcode for D hemisphaericum

from Philippines and Vietnam populations

◉ to construct phylogenetic tree D hemisphaericum coming from two neighboring

countries Vietnam and Philippines

◉ to analyze the genetic diversity of a clear cut morphospecies, D hemisphaericum

1.4 Scope and Limitations

These research frameworks were conducted to (i) predict habitat suitability of

Diderma hemisphaericum in the current and future scenarios and (ii) assess the genetic

diversity of D hemisphaericum, a cosmopolitan myxomycetes scpecies The whole study

was conducted from June 2018 to September 2018 in the Institute of Botany and Landscape Ecology, University of Greifswald, Greifswald Germany

However, some limitations of this study includes: (1) the species occurrences

coordinates where Diderma hemisphaericum were found, are mostly located on the Northern sland of the Philippines (Luzon) Thus, fewer predictive tests have been carried out in the Southern part (Mindanao) region which resulted in a lower probability in that

area (2) Only the 19 bioclimatic variables have been used as an environmental layer for

D hemisphaericum, which may also be affected by certain variable like elevation; and (3)

there are only 46 species occurrences which could affect the species distribution model (4) For DNA barcoding, there are limited numbers of specimen in Vietnam; (5) most of

the specimens in Philippines were collected in the Luzon Island and lastly, (6) the amount

of allocated time and money were limited for doing two research frameworks

1.5 Definition of terms

For the sole purpose of understanding, deep and scientifically specialized words here are the list of terms used for this study and their corresponding meaning

18srDNA barcoding refers to the DNA barcoding process which uses SSU gene marker

A2 climate scenario refers to the future climate scenario in which the earth’s technology

will remain as is, thus, resulting in high increase of temperature

Allopatric speciation refers to the change of a species from its ancestral species due to

geographical barriers

B1 climate scenario refers to the future climate scenario in which the earth’s technology

will become more efficient, thus resulting in a low increase of temperature

Bioclimatic variables refers to the 19 abiotic variables that may affect the species

distribution

Biospecies refers to the individuals in a certain species that are capable of sexual

reproduction

Diversity refers to the variety of species and usually measured by richness and evenness

DNA amplification refers to the process of artificially creating more DNA fragment

copies through replication with specific primer pairs

DNA sequence refers to the results of the DNA barcoding process that can be viewed and

Fruiting body refers to the life stage of myxomycetes as a plant that release spores in the

air for reproduction

Gene flow refers to the transfer of the genetic variation from one group to another

Gene marker refers to the DNA sequence that has a known location for identifying

individuals or species

Gene network refers to the visualization of the evolutionary relationships between

sequences

Genetic diversity refers to the variety of genes within species

Habitat suitability refers to when organisms can successfully reproduce and grow in an

area or condition

Maximum entropy refers to algorithm that uses a least-biased statistical interference

when insufficient information is available

Maximum likelihood refers to the statistical method that randomly draws a given sample

from a population over the possible values of the population parameters

Morphospecies refers to the indivuals that are named based on their morphological

properties

Niche refers to the environment where a species or organisms can successfully grow and

reproduce

Population refers to the number of a species or organisms that occupies that same area

Phylogenetic tree refers to the diagram that shows the evolutionary relationships among

different biological species

Ribotype refers to to RNA component of the genetic material of a species or organisms

Species distribution modelling refers to the process where the species occurrence

datasets and environmental layers are combined and modeled using different algorithms

to create a predictive map of distribution

Statistical parsimony refers to the statistical approach that minimizes the total number

character-state changes

Sympatric speciation refers to the change of a species from its ancestral species in a

same geographical region

lianas (DeWalt et al., 2015), spiders (Boyer et al., 2016) and mammals (Bennie et al., 2014), but limited models proposed for microorganisms Nonetheless, understanding the most appropriate biogeographical hypothesis of global distribution is important particularly on the light of the undergoing estimation of the Earth’s total biodiversity (Gonzalez et al., 2016; McGill et al., 2015) For example in the case of myxomycetes, we can assume that if all myxomycetes are widely distributed, then their global diversity is low On the other hand, if myxomycetes are restricted in terms of their geographic distribution and their patterns are not by merely caused by their expected habitat requirements, then their global diversity should be high (Mitchell & Meisterfeld, 2005) In the next section of this review, two opposing hypothesis are discussed to present the current understanding about the less explored biogeography of myxomycetes

The first hypothesis is known as the Everything is Everywhere (EiE) model

According to this model, due to their relatively small spore size, like other microorganisms, myxomycetes or true plasmodial slime molds are expected to follow the

Baas-Becking model of microbial ubiquity in nature (Finlay, 2002; Finlay & Clarke 1999)

to actually grow and develop For example, the Taimyr Peninsula, located in the Far

North of Russia, has an extremely low temperature but Cribaria violacea, a type of

myxomycetes that can be found usually in tropical regions, has been found in this region (Novozhilov et al., 1999) Another example is the recent new records of Redeña- Santos

et al., (2018) where they found that what seems to be exclusive temperate myxomycete are reported for the first time in the subtropical climate of Northern Vietnam Similar

findings of Diderma hemisphaericum, a myxomycete abundant in temperate region has

been spotted in the tropical climate of the Philippines (see studies of Dagamac et al., 2015b & c; dela Cruz et al., 2010) This model is further supported by other studies that showcased rapid rates of dispersal for some metacommunities of protists (see studies of Berdjeb et al., 2018; Warren, 1996) Most of these studies showed that protist are capable

of long distance dispersals due to several abiotic factors i.e wind (van der Gast, 2015), volcanic eruption (Bass & Boegnik, 2011) or the intervention of biological vectors of dispersal such as birds (Wilkinson et al., 2012), and even human movements (Perigo et al., 2012)

However, an opposing second hypothesis about these general biogeographical patterns for protist had also arises A popular model that refutes the former model is the

model first proposed by Foissner (1999) which is known as Moderate endemicity (ME)

According to this model, even if there are myxomycetes species that can be cosmopolitan

or widely distributed globally, portion of the ca 1000 morphospecies (Lado, 2018) have

geographical barriers i.e mountain ranges (Boenigk et al., 2016), oceans (Dagamac et al., 2017a) may serve as a hindrance for myxomycetes to disperse and be distributed like in

the case of plants (Ixora, Mouly & Pisivin, 2007; Nepenthes, Van der Ent et al., 2015;

Cinnamomum, Sriramavaratharajan et al., 2016) or animals (Tropidurus tropiduridae,

Carvalho et al 2018; Mimus parvulus, Gotanda et al 2015; Drosophila, Caddock et al.,

2016), that are argued to have a native restriction at a local scale (Coterill et al., 2008) In the arguments published by Foissner (2008), the ME model contradicts EiE by proposing the following parameters: (i) migration rates are lower in 90% of the known species therefore yielding at the same time low abundance rate, (ii) the ability of a species to get extinct will be moderate, (iii) the ratio of the global species pool that are found in local scale setting is again moderate and (iv) at least 30% of the morphospecies are considered

to be restricted or endemic In cases of recent myxomycete studies, it seems that this model is more strongly favoured Like for an instance is the case of the genus

Ceratiomyxa that showed two species namely Ceratiomyxa morchella and Ceratiomyxa sphaerosperma (Rojas et al., 2008) are found only in the tropical region of Central

America Moreover, patchy distributional range of the unfathomable myxomycete species

Barbeyella minutissima (Schnittler et al., 2000) points out the exclusivity of this

myxomycete species to temperate Picea and/or Abies type of forests However, these

observations are just only recently supported by Dagamac et al 2017c using sophisticated ecological community models that compared the diversities and distribution of tropical myxomycetes In that study, a clear bifurcation between the Neotropical and Asian Palaeotropical forest myxomycetes communities was illustrated and this was caused by

the number of regional endemic species recorded for the two forests In addition, recent molecular data of several myxomycete species that studied populations in genetic levels also supported the moderate endemicity hypothesis and rejected the EiE model As such, the two comprehensive genetic studies using myxomycetes models were the studies of

Aguilar et al., 2015 for Badhamia melanospora and Dagamac et al., 2017a for

Hemitrichia serpula Interestingly, both of the population genetic studies employed two

important methods to support such suppositions for moderate endemicity These methods are the DNA barcoding technology and the species distribution modelling which are the current trends for biogeographical studies

2.2 SPECIES DISTRIBUTION MODELING (SDM)

According to Guisan and Thuiller (2005), a vast number of modeling techniques that predicts the distribution of a particular species has been recognize nowadays as an important element of conservation strategies and environmental management purposes Most of these modeling techniques are found to be comprehensible since it only need two important components namely (i) the occurrence data or the dataset where the species are reported to be present and (ii) environmental layers Models then utilize the associations between the two components to give an impression about the most probable environmental conditions where the populations of that particular species can be retained There are five important elements of SDM research according to Zimmerman et al (2010) namely, (1) the biotic interactions, (2) niche stability, (3) evolutionary processes, (4)

importance of design for sampling and (5) the species invasion One of the most common approaches is distinguishing the environmental factors in which the species are suitable to flourish and then recognizing where habitable environments are distributed to estimate the geographical distribution There are two principles of environmental conditions used for modeling: (i) mechanistic and (ii) correlative The former model has a more complicated understanding because it requires the knowledge of physiological response of a particular species relative to the environmental factors (e.g Kearney & Porter, 2009) while the latter model, can yield more productive results because only the known occurrence records combined with the environmental variables are used to determine the effects on the rate species persistence (Rode & Lieberman, 2005; Graham et al., 2004) For this type of modeling, data on different environmental variables such as (precipitation, temperature, etc.) and the species records will go through an algorithm (i.e Maximum entropy, Gower method, Artificial Neutral Network and the like) to find which among the variables that

was tested has the most influence on the species For example, Thuja sutchnuencis, an

extremely endangered conifer species in Southern China, was modeled using MaxEnt algorithm and associated that among the 19 bioclimatic variables found in the world climate database, the precipitation of the warmest quarter of the year was the most relevant factor (Qin et al., 2017)

But perhaps, the application of species distribution modeling provides many advantages Like for an instance, in finding populations of a known species (Pearson et al., 2007; Guisan et al., 2006), a study of Wasof et al (2015) revealed that SDM can

effectively address climatic niches in disjunct population of European vascular plant species Moreover, SDM can also help in assessing the impacts of climate change (Pearson & Dawson, 2003) An example of this is a study of the mountain pine beetle,

Dendroctonus ponderosae, in the Rocky Mountain, United States using different climate

change scenarios to compare the current and predicted geographical distribution (Dowling, 2015) Furthermore, the problem about invasive species can also be addressed using SDM methods (Václavík & Meentemever, 2009; Wiens et al., 2009; Peterson,

2003) For instance, Bromus tectorum, an invasive cheat grass in the Rocky Mountain

National Park Colorado, USA, has been validated and tested using the maximum entropy algorithm (West et al., 2016) In addition, the application of this model is widely utilized

in many conservation studies (Costa et al., 2010; Pearson, 2007) An endangered bird

species in Taiwan called Terpsiphone atrocaudata has a threatening situation regarding

their population decline and by using SDM, several factors can be determined to help preserve and maintain this species (Spath et al., 2016)

In spite of these promising applications of SDM, very limited studies showed its utilization among microbial groups So far, studies in lichens (see Dymytrova et al., 2016; Braidwood & Ellis, 2012), ectomycorrhizal fungi (see Guo et al., 2017; Wolfe et al., 2010), protosteloid amoebae (Aguilar & Lado, 2012) or insect viruses (see Larson et al., 2010) have applied MaxEnt algorithm to address some ecological questions In the case of the myxomycetes, only few papers are so far published that used SDM Rojas et al (2015) used this modeling technique to (1) get a better understanding with the most commonly

occurring myxomycete species in Costa Rica (2) and establish future baseline assessments

of myxomycetes in relation to the changing climate To further prove the speculation about patterns of global distribution among genetic clades in myxomycetes, SDM was

also performed in species of Badhamia melanospora (Aguilar et al., 2015) and

Hemitrichia serpula (Dagamac et al., 2017a) to support the aforementioned moderate

& Francis, 2012) sequences which acts like as an identifier to a particular species (Tautz

et al., 2003) The intention was very clear at the beginning wherein a comprehensive database of all biological sequences, most preferably assigned with a known voucher specimen that corresponds to a described species, is matched to sequences from sampled individuals (Blaxter, 2004) This now revolutionized many applications and possibilities

in many areas of research For example in the field of agriculture and forestry, it is used to analyse the phylogenetic relationships of trees in tropical forests (Kang et al., 2017) In the field of forensics, DNA barcoding has been used to indentify illegal trading of African

Pangolin scales (Mwale et al., 2016) There is also a study about using this method in plants to be a key factor in solving a crime e.g [i] movement of corpse across the field, [ii] route checking of an identified suspect, and determining narcotics (Park et al., 2017) Another application of DNA barcoding is the fraud surveillance in the market Like for

the case of many processed foods (processed tuna, Nurilmalaa et al., 2016; fishballs,

Shokralla et al., 2015) or herbal medicines (Gao et al., 2017; de Boer et al., 2015) where many of the ingredients that are claimed to be in the products are not really there or was replaced by other species constituent Moreover, in the field of conservation strategies, DNA barcoding was employed in many harmful organisms that poses high biological risks in globally competitive economic market i.e anthropod pests (Floyd et al., 2010)

Nevertheless, these promising applications of DNA barcoding are harnessed with one very precarious point: the standardization of selecting the appropriate part or segment

of a genome (from this point forward is addressed as a gene marker) that can tell different species apart (Taylor et al., 2012) This is because the reliability of the technology relies

on whether or not the variation in the sequences can really differentiate species As a rule

of the thumb, the ideal DNA barcode should at least showcase (i) a high interspecific variability (between different species) and (ii) a low intraspecific variability (within the same species in a population) concurrently This leads to different gene marker candidates

across different biological species in the tree of life For the animals, cytochrome c

oxidase subunit (COI) has proved to be 95% successful barcode marker (Hebert et al.,

2016) This has been used as a barcode to animal species like insects (Kranzfelder et al.,

2017; Karthika et al., 2016; Jalali et al., 2015), birds (Bilgin et al., 2016; Aliabadian et al., 2013; Hebert et al., 2004) or even amphibians (Lyra et al., 2017; Dang et al., 2016; Vences et al., 2012) Conversely, plant barcoding poses some challenges due to the fact that some plants species have intricate breeding systems such as hybridization (Fazekas et al., 2009), introgression (Liu et al., 2016) and polyploidy (Braukmann et al., 2017) but

perhaps what is widely used as the genetic barcodes for plants species are either matK (maturase K) and/or rbcL (ribulose biphosphate carboxylase) For fungi, the internal

transcribed spacers or ITS region (Begerow et al., 2010, Nilsson et al., 2008) shows discriminatory capability as a genetic barcode in some fungal species like in the case of Agaricomycotina (Dentinger et al., 2011) or the genus Hypoxylon (Suwannasai et al., 2013) With regards to the protist group, the 18S rDNA seems to give the best resolution

so far (see studies of Santos et al., 2017; Decelle et al., 2014) Even though, the investigations that are conducted for the protist group are still relatively new, the usage of this gene marker as a universal barcode for the group has able to provide at least the discrimination up to the phylum level (Hebert et al., 2016) Hence, other gene markers are explored to test species discrimination with better resolution in taxonomic placement For example COI is used for oomycetes (Robideau et al., 2011; Levesque, 2011) or nebelid testate amoeba (Kosakyan et al., 2012) and rbcL is suggested to be appropriate for the diatoms (Hamsher et al., 2011) But since the protist lineage is very diverse, finding the most suitable genetic marker that would serve as a barcode across the whole kingdom would be an arduous task to accomplish, especially for myxomycetes whose molecular era started quite late

In the case of the myxomycetes, a recent review about the different genetic markers applied in myxomycetes was published by Schnittler et al (2017) On their review, different gene markers (i) SSU 18sRDNA Small ribosomal unit (ii) ITS: Internal Transcribed Spacer (iii) EF1A: Elongation factor 1 alpha gene (iv) COI: Mitochondrial cytochrome c oxidase subunit I gene and lastly (v) mtSSU: 16s rRNA mitochondrial SSU gene, have been discussed and tested to know which would be the ideal barcoding gene marker for the myxomycetes Of all the said gene markers, the SSU gene marker has the

most promising results because the sequence is almost free from non-coding regions

(introns) that may disrupt some sequences of similar species Moreover, the SSU gene marker was predominantly used in resolving deep phylogenetic (how related a species from another species) relationships in many myxomycete studies (Fiore- Donno et al., 2008; Fiore- Donno et al., 2005) In more recent studies about myxomycetes barcoding, SSU was utilized in environmental samples like soil (Borg-Dahl et al., 2018), dead wood (Clissman et al., 2015; Ko Ko et al., 2009) and or even in the multinucleated plasmodial stage of the myxomycetes (Shchepin et al., 2017) Nevertheless, more studies and experiments are needed to further verify if SSU is the most accurate universal gene marker for myxomycetes (Schnittler et al., 2017)

PART III

METHODOLOGY

3.1 Predicting local habitat suitability in changing climate scenarios

3.1.1 Gathering the distribution records of D hemisphaericum in the Philippines

Initially, all published literatures about the distributions and diversity of Philippine myxomycetes were surveyed From these studies, only those that (i) reported the presence

of Diderma hemisphaericum and (ii) annotated a valid coordinate of the locality were D

hemispahericum was found were collected for the analysis Overall, 46 coordinates that

recorded the occurrences of Diderma hemisphaericum from the Philippines was used for

this research framework The valid coordinates were considered now as a presence- only input file (species occurrence data) with negligible sampling bias with respect to sampling conditions The input files were converted into geographic coordinate system (latitude and longitude) format and was then saved as a comma delimited or CSV file type

3.1.2 Obtaining current and future (A2 and B1) environmental layers

In this study, two different environmental scenarios, which in this case, the 19 bioclimatic variables, in the Philippines will be used: (i) the current climate (for the period

of 1950- 2000) at 2.5’ resolution, which can be obtained in the PhilGIS web portal (http://ww.philgis.org) (see also Hijmans et al., 2005), and (ii) A2 and B1 climate scenarios for the year of 2080, which can be also downloaded in (http://www.ccafs- climate.org) following Banag et al (2015) protocol These two future scenarios (A2 and

B1) were chosen because they represent two different storylines of climate change The A2 scenario assumes a continuously growing population and higher emission rate with a global mean temperature forecasted to rise by 3.4° Celsius On the contrary, the B1 scenario assumes a change in material use and technology advances in a more efficient way hence; it predicts that the global mean temperature to increase by only 1.8° Celsius These two future scenarios were downloaded as an ESRI file and were then converted to ASCII format to match the requirements in the MaxEnt software

3.1.3 Species distribution modeling

The species occurrence data that was transformed as a presence- only input file of

Diderma hemisphaericum and (19) bioclimatic variable for the local region of the

Philippines were modeled using MaxEnt v 3.4.1 MaxEnt is an open- source software that

is used for modeling species niche and distribution that utilizes machine- learning based

on maximum entropy algorithm (Philips et al., 2004) Typically, MaxEnt software is applied for conservation purposes (Yang et al., 2013; Kumal & Stohlgren, 2009; Thorn et al., 2009) because it provides more accurate predictions than any other species distribution modeling algorithms (Townsend- Peterson et al., 2007) Furthermore, MaxEnt create better predictions even though the sample size (species occurrence data) is relatively small (Costa et al., 2010) Here, the two datasets (1) species occurrence data and (2) bioclimatic variables were subjected in the MaxEnt software using the default settings

Furthermore, the option “Create a response curve” and “Do jackknife test to measure

variable importance” were ticked and then the output format was set to “logistic” in order

to evaluate the importance of each biophysical variable The output file that was generated from MaxEnt was then exported as ASCII file format To get a detailed visualization, the

ASCII output file format from MaxEnt was imported in ArcMap 10.4 software In the

ArcMap software, the map was then divided using defined interval as a classification

method into four different categories (i) very low probability (VLP) which values range from 0- 0.25 (ii) low probability (LP) which range from 0.25- 0.50 (iii) medium probability (MP) which values range from 0.50- 0.75 and lastly, (iv) high probability (HP) which values range from 0.75- 1.00

Figure 2 A diagram showing major steps for constructing a robust correlative species distribution model

3.2 Assessing genetic diversity across regional populations

3.2.1 Obtaining samples of Diderma hemisphaeriucm

A total of 50 samples of Diderma hemisphaericum were obtained from the

herbarium collection of the University of Greifswald From this collection, two set of Southeast Asian populations namely from the (i) Philippines and (ii) Vietnam were selected for this study

3.2.2 DNA extraction, amplification and sequencincing

Following the protocol of Dagamac et al (2017a), approximately 10 mature fructifications coming from the herbarium collections were manually transferred into a 2

ml sterile safe- lock eppendorf tubes with glass beads and coarse sand To maintain the aseptic condition during the transfer of fruiting bodies, the working space was initially washed with 70% EtOH To further avoid cross contamination between the samples, the forceps were chemically sterilized by washing it concurrently with 1M HCl, detergent solution and 70%EtOH After transferring, the samples were then immediately mechanically homogenized using a vortex (Bead Blaster™ 24) Next, the DNA extraction using Mag- Bind Plant DNA Kit (Omega Bio- tek, Georgia, USA) was done following the manufacturer’s protocol After, the processing of the magnetic bead set- up, the samples were then extracted using the robot model KingFisher™ Flex Then, the amplification process of the extracted DNA samples has been carried out by mainly making the following combination for a master mix PCR solution: 5.64 µL of H20, 0.50 µL of

colored buffer (5x Mgo-Buffer), 1µL of MgCl2 (Bioline, 50 mM), 0.20 uL of dNTPs, 0.25 of forward and reverse primer (S1 and Su19Rsp), 0.16 µL of Mango- Taq DNA Polymerase (5U/µL) and lastly, 2 µL of diluted DNA was added totaling in10 µL in volume Next, to check if the amplification process was successful, visualization of the DNA was performed by using agarose gel electrophoresis For a prepration of 1.8% of agarose solution, 3.6 g of agarose were mixed with 200mL of TAE buffer After preparing for the agarose gel, 4µL of each PCR product was added with 4 µL colorng dye on a parafilm The mixture was then carefully displaced inside the open wells of the agarose chamber After all samples were placed, the electrical machine was turned on for approximately 45 minutes at 100V Afterwards, the gel was placed inside an ethidium bromide solution for ca 30 minutes Then, it was dipped for 10 minutes in a water bath to wash the toxic ethidium bromide solution on the gel To view the generated bands on the gel product, UV trans-illuminator (BST- 20G- 8R) was used All successfully amplified DNA samples that have produced clear bands were separated for the purification step using enzymes (Exonuclease and alcalic Phosphatase - EAP) After subjecting the purified amplicons (37º Celsius, for 1 hour and 85º Celsius for 15 minutes), the concentration was measured using Nanodrop (Nanodrop™ Lite Spektralphotometer Thermo Scientific™) It was then adjusted to get the final concentration of 10 ng/µL

Now to perform the cycle sequencing process, the following mastermix combination was prepared: 1.5 µL sequencing buffer, 4.69 ddH2O, 4.00 µL big Dye, 4.00

µL Half big dye and 0.625 µL of the S1 forward and reverse primer Using forward and

reverse primers for DNA analysis can be useful for checking the reliability of each sequence Then, the products of the PCR process were cleaned with serial ethanol washing (30 µl of 96% EtOH and 100 µl of 70% EtOH) After overnight incubation, the samples were then placed inside a 96 well plate were sent to the Zoology department of the University of Greifswald for sequencing purposes using the ABI3730 sequencer

3.2.3 Quality filtering and sequence alignment

The qualities of the DNA sequences were initially validated using the sequence

reader software Chromas and were cross checked using the Nucleotide BLAST (Basic

Local Alignment Search Tool) Database Sequences that gave an E-value of 0.00 and has

a high percentage similarity to Diderma sequences (for this case, only valid publications

that reported the corresponding sequence) found in the database were then used to the subsequent analysis After quality filtering of the generated sequences, an automatic

alignment algorithm using the online version of MAFFT (multiple alignment using fast

Fourier transform) software were utilized In this online software, the following

parameters were used: E- INS- i setting with a 2.5 gap penalty The new alignment file

was then manually checked for 505 base sites for consistency using BioEdit v7.0.5 and

those that have inconsistent bases were validated with the sequence quality

3.2.4 Phylogenetic tree and ribotype networking

Continuing, a phylogenetic tree was constructed using the software MEGA v7.0

(Molecular Evolutionary Genetics Analysis) based on the Maximum Likelihood algorithm with 1000 bootstrap replication and GTR=Gamma for nucleotide substitution An

outgroup, Physarum melleum (Genbank accession no: KC759095), was also added to

serve as the root of the phylogenetic tree The tree was then observed and edited using the software A second analysis based on the most statistically parsimonious gene network for ribotypes having 95% coherence in accordance to the set of possible outcomes based on coalescent theory (Hudson, 1989) was also constructed for this study The ribotype networks separated by mutational steps were created using the software TCS v.1.21 (Clement et al., 2000) which is based on the calculation of the probability that pairs of ribotypes are similar for all other combinations of ribotypes and then connects only the

most similar ribotypes together into a network where their combined probability is >95%

Therefore, the resulting network will remove divergent ribotypes whose true genealogy may be concealed by homoplastic characters (Templeton and Sing, 1993) For better visualization of the phylogenetic tree and the ribotype network, both figures were reconstructed manually

3.2.5 Analysis of the genetic diversity

To estimate the genetic diversity of D hemisphaericum population in (i)

Philippines and (ii) Vietnam, a RAC (ribotype accumation curve) were constructed to calculate in EstimateS v9 (Colwell, 2013) with 200 randomization values running on

“default settings” Next, the correlation between the genetic and geographic distances was also calculated by constructing two independent matrices; (i) matrix consistsing of the genetic distance from the 505 bp multiple sequence alignment of samples applying the maximum composite likelihood model implemented in MEGA software and (ii) the matrix which includes the geographical distances between the collection sites of any two samples being compared as calculated using an Excel template with the Vincenty formula (Dagleish, 2015) The the two matrices were then subjected to a Mantel test with 999

iterations in the ExtraStats function in PopTools v3.2.5 (Hood, 2010) Afterwards, to

detect the differentiation of ribotypes between the Philippine and Vietnam populations, an AMOVA (analysis of molecular variance) and genetic differentiation calculated as Chi-

square statistics was done using Arlequin 3.5.2.2 (Excoffier & Lischer, 2010) with a

significance based on 10000 permutations To analyze the polymorphic sites that yielded the diversity of the ribotypes, nucleorides and mean number of nucleotide differences were implemented using the software DnaSP v 5.10 (Librado & Rojas, 2009)

The maximum entropy algorithm used for species distribution modelling for

Diderma hemisphaericum suggested different important bioclimatic variables that

probably influence the predictive local distribution of this species in the Philippines Taking into consideration the measure of percent contribution, both for the current predictive distribution and A2 climate scenario, isothermality has the highest percentage contribution with 41.3, and 29.3, respectively, but for the B1 climate scenario the highest contribution was attributed to temperature seasonality (23.6%) In terms of permutation importance, the model calculation indicated isothermality to be highest in the current scenario with 47.7 and temperature seasonality for both the future climate change scenarios with 41.2 for the A2 scenario and 28.9 for the B1 scenario Moreover, the Area Under the Curve (AUC) values from the maximum entropy-based results for the different climatic scenarios are (i) 0.80 for the current climate scenario; (ii) 0.90 for the A2 climate scenario; and (iii) 0.88 for B2 climate scenario which point to a good quality of predictive model performance (Table 1) Furthermore, a Jackknife test revealed that the predictor temperature seasonality contained the most information among the 19 bioclimatic variables for creating the models for the current distribution and two changing climate scenarios The predictive model (Fig 3) based on the current species occurrence data