RAPD and

ISSR analysis details

In present study, initially 40

RAPD primers that is 2 set of Operon primer kits OPG and OPR (20 primer from

each kits) were used to detect genetic polymorphism of S. oblonga, S. fruticosa, S. chinensis and S. macrosperma. Out of the 40 RAPD primers, 10 primers i.e. OPG-02,

14, -16, -17, -18, -19 and OPR-02, -03, -07, -08 showed reproducible amplified

DNA polymorphism. All the chosen primers amplified fragments across the 21

samples, with the number of amplified fragments ranging from 4 to 12. Minimum number

of loci were seen in the primer OPG18 (4 bands) and maximum bands were observed

in primer OPG17-12 bands. From the ten primers, a total of 76 loci were

generated of which 70 were polymorphic, making polymorphism generated by RAPD

makers to be 92.11%. Multiplex ratio of RAPD analysis was calculated to be 7.6.

Cumulative resolving power of 10 RAPD primer was 54.67.

While in ISSR analysis 10 primers

produced 67 loci of which 61 bands were polymorphic, accounting for 91.04% of

polymorphism. Number of loci varied from minimum of four in primer ISSR 5 to

maximum of nine in ISSR 10. Multiplex ratio of ISSR analysis is calculated to

be 6.8. Cumulative resolving power of 10 ISSR primer was 58.48. The marker

index for RAPD and ISSR was 6.54 and 5.45 respectively.

Observed number of alleles,

effective number of alleles, Nei’s genetic diversity, Shannon’s information

index, for 21 samples of Salacia

species analyzed using ten each of RAPD

and ISSR primer were found to be 1.9211, 1.4537, 0.2785, 0.4294 and 1.9104,

1.5108, 0.2988, 0.4509 respectively. Total genotype diversity among population (Ht) was estimated to be 0.2713 while within population

diversity (Hs) was estimated to be 0.1514 for RAPD and for ISSR Ht was 0.3055

and Hs was 0.1222. Mean coefficient of gene differentiation (Gst) value for

RAPD was 0.4418 and ISSR was 0.5999. Suggesting that 55.8% and 40.1 % of the

genetic diversity resided within the population as per RAPD and ISSR markers.

Estimates of gene flow in the population for RAPD and ISSR were 0.6318 and

0.3334 respectively. (Table 3).

Dendrogram and PCoA of RAPD and ISSR

In RAPD dendrogram, 21

samples of Salacia grouped into two clusters

(Cluster 1 and 2). Cluster 1 contained S.

chinensis SC1 to SC5 and cluster 2 was further divided into two

sub-clusters (sub-cluster 1 & 2). In cluster 2, sub-cluster 1 contained all

samples of S. macrosperma along with

two samples of S. fruticosa SF1 &

SF3 (Fig 1) and sub-cluster 2 contained three remaining samples of S. fruticosa along with S. oblonga samples. The cumulative total

variation of three principle components accounted for 65.68 % of variation.

Dendrogram of ISSR data

showed that the samples clearly grouped into four clusters (I, II III and IV)

of its respective species S. chinensis,

S. macrosperma, S. fruticosa, S. oblonga. For ISSR analysis cumulative

total variation of three principle components accounted for 74.05% of the

variation. The results of RAPD and ISSR PCoA analysis were comparable to the

cluster analysis (Fig 2).

ITS analysis

For ITS analysis, 19 samples of

current study and an outgroup Pristimera

preussii belonging to sub-family Hippocrateoideae was used to construct

phylogenetic tree. Two samples SM1 and SM14 produced faint bands and could not

be sequenced. Sequence alignment of 20 samples resulted in overall sequence

length of 752 bp, of which 221 bp (29.38%) were conserved, 503 bp (66.88%) were

variable sites and 103 bp (13.69%) were parsimony informative sites. Three

major clades were observed from ML tree. Clade 1 contained all the samples of S. macrosperma along with samples of S. oblonga which were nested with-in the

clade. Clade 2 and 3 contained S.

chinensis and S. fruticosa

samples respectively (Fig 3).

Comparative analysis of population

Values of observed number of

alleles, effective number of alleles, Nei’s genetic diversity, Shannon’s

information index of each population were compared to observed diversity and

degree of polymorphism with-in the population (Table 3 and 4). In comparison,

the RAPD values were marginally higher than the ISSR except in the S. fruticosa population. Significant

differences were observed in all the parameters. Highest percentage of

polymorphism and highest polymorphic loci were seen in S. fruticosa population in RAPD analysis. In RAPD, ISSR and ITS analysis

high degree of polymorphism was seen in S.

macrosperma and S. fruticosa population

followed by S. chinensis population. Although

only two samples are in S. oblonga

population, RAPD, ISSR and ITS analysis detected polymorphism of 15.79%,16.42%,23.36%

respectively. Also, in parameters such as Ht, Hs, Gst and Nm significant

differences in value were observed (Table 4). The level of polymorphism

revealed by RAPD was (41.45%±10%) which was higher than ISSR (33.58%±6.52%) and

ITS (25.50%±17.25%). The polymorphism of each population of S. chinensis, S. macrosperma, S. fruticosa and S. oblonga from RAPD was35.53%,

55.26%, 59.21%, 15.79%and ISSR was32.84%,

47.76%,37.31%,16.42% respectively.

For comparative analysis of ITS

with the RAPD and ISSR, sequences data of ITS was analyzed in GenAlEx. Before

exporting the data, the outgroup sequence and sequence SF5 were removed. Only

polymorphic nucleotide positions were converted to numeric codes (A=1, C=2,

G=3, T=4, hypen/colon=5) and 137 sites showed the polymorphism which were used

for the further analysis. The polymorphism of each population of S. chinensis, S. fruticosa, S. macrosperma

and S. oblonga from ITS analysis was

6.57%,19.71%,24.82%,23.36% and overall polymorphism was 18.61%±4.16%. For ITS coefficient

of evolutionary differentiation was 0.797which indicated that 20.3% of the

genetic diversity resided within the population. Tajima’s D neutrality tests were

performed to check whether genus Salacia

populations followed a neutral model of evolution with constant population size

over time. The observed values of Tajima’s D neutrality tests were -1.089757

for S. macroperma and S. oblonga population, -1.105205 for S. fruticosa and -0.174749 for S. chinensis and-1. 181277 for all the

19 samples. After removing sample SF5 since it showed high divergence,

neutrality test was performed for 18 sample of Salacia which gave observed value of 0.606285.

AMOVA, which helps in

partitioning of the overall variations among groups and among populations

within the group were performed for RAPD, ISSR and ITS data matrices. From RAPD,

39% of molecular variance was found among population while, within the population

this value was found to be 61% indicating that there were more variations

within the population. While in ISSR, 55% molecular variance was found among

population and 45%within the population. For ITS sequence analysis 80%

variances was among the population and 20% variance was within population which

was similar to coefficient of evolutionary differentiation. (Table5).

Nei genetic pairwise

distance of Salacia species was found

to be > 0.5 for RAPD, ISSR and ITS sequence. But in ITS sequence analysis,

the pair-wise distance between the S. oblonga

and S. macrosperm was 0.061

suggesting that they are very closely related. In addition, the pair-wise

distance and identity of S. oblonga

and S. fruticosa was 0.915 and 0.088

indicating that they are highly dissimilar. (Table 6).

Statistical comparative analysis

Mantel test was employed to

determine the coefficient of correlation between the genetic distance matrices

generated by RAPD and ISSR markers. The coefficient of correlation between RAPD

and ISSR marker was R2=0.3781, r=0.614 which is high. This value signifies

that there was considerable correlation between RAPD and ISSR genetic distances

matrices. Twenty-one samples grouped into two clusters in RAPD dendrogram

whereas in ISSR dendrogram four cluster were observed. Comparing RAPD

dendrogram with ISSR dendrograms we can notice that S. oblonga was an Operational Taxonomic Units (OTU). In all

analysis, results of cluster analysis were comparable to PCoA.

Mantel test was also employed to

analyze the ‘goodness of fit’ for each marker system. This was done by

comparing cophenetic similarity matrices of genetic distance with cophenetic

similarity matrices with the Nei’s Genetic Distance for each marker technique.

It revealed values higher than 0.80 for all the markers used RAPD (r = 0.827,

P = 0.01), ISSR (r = 0.816, P = 0.01) thus confirming their authenticity and

very good fit of PCA clustering.