Background
The key to the toxicity of Bacillus thuringiensis against the insect larvae is the specific molecular interactions of the insecticidal proteins with the membrane receptors followed by pore formation in the insect mid-gut epithelium. The degree and spectrum of toxicity of Bt insecticidal proteins against different insect species are variable. There are currently around 75 primary subgroups of Cry toxins, 3 for Cyt toxins and 4 for Vip toxins (Adang et al., 2014) and more than 300 different members so far reported are present in these subgroups (Crickmore et al., 2014, http://www.btnomenclature.info/). The remarkable diversity is because of a high degree of genetic plasticity or variations that occurs among the Bt strains due to many intrinsic factors like the presence of many different plasmids in each strain and their conjugal transfer, recombination between chromosomal DNA and plasmids, involvement of transposon-like inverted repeats flanking the endotoxin genes in high frequency causing DNA rearrangements etc and some extrinsic factors like mutation, nutritional influences etc (Kaur et al., 2006). Genetic diversity among the Bt strains is, therefore, beneficent which enhances the scope of discovering novel toxins and urges for extensive exploration for more strains.

Again, resistance development in insects against any insecticide is a common occurrence with no exception for Bt toxins. The facts behind the resistance and cross-resistance of insect pests to Bt toxins as reported are, i) reduction of binding of toxins to receptors in the midgut of insects, ii) reduced solubilisation of protoxin, iii) alteration of proteolytic processing of protoxins and iv) toxin degradation and or precipitation by proteases etc (Bruce et al., 2007). Few additional virulence factors such as, phospholipase C (Palvannan and Boopathy 2005; Martin et al., 2010), proteases (Hajaij-Ellouze et al., 2006; Brar et al., 2009; Infante et al., 2010) and hemolysins (Gominet et al., 2001; Nisnevitch et al., 2010) etc that influence the toxicity, are also variable among the strains due to the genetic variations beside the insecticidal proteins.

Efforts should, therefore, be continued in discovering more Bt strains with genetic diversities to discover novel toxins with improved activity, to widen the spectrum and to overcome the resistance problems besides protein engineering with the existing pool. So, a good number of Bt strains were previously isolated by us from different regions of Bangladesh and studied in terms of their abundances, distribution patterns as well as diversities connected to biochemical properties, plasmid and cry genes profile (Shishir et al., 2014). Potentiality of these strains, presumed through insecticidal gene and protein profile analysis, was established for certain strains against Melon fruit fly (Bactrocera cucurbitae) upon bioassay and this novel toxicity was reported (Shishir et al., 2015).

Thus, more Bt strains and the genetic diversity analysis among them is highly significant for maximum utilization of the resources and to combat the resistance problem. Several different techniques were reported for genetic diversity analysis, like M13 fingerprinting (Miteva et al., 1991), arbitrary primer PCR (Brousseau et al., 1993), PCR using conserved primers for 16S to 23S ribosomal intergenic spacer sequences (Bourque et al., 1995), DNA hybridization using variable region of 16S rRNA gene (te Giffel et al., 1997), AFLP fingerprinting, RAPD-PCR (Welsh and McClelland, 1990) etc. Ribotyping, either by PCR or DNA hybridization, failed to detect the diversity among Bt strains which could be because of the use of one single gene or operon and the evolutionarily conserved nature of rRNA gene. However, when the whole genome was used for identification of Bt strains by M13 DNA fingerprinting and arbitrarily primed PCR, considerable diversity among the Bt serovars representing different serotypes had been detected. Random amplification of polymorphic DNA (RAPD) is a modified method of PCR with a single arbitrary primer that recognizes differences in the prevalence and positions of annealing sites in the genome and produces a varied spectrum of amplicons reflecting the genomic composition of the strains (Welsh and McClelland, 1990; Williams et al., 1990). Besides this taxon-specific amplicon generation for the detection of diversity, RAPD-PCR fingerprinting technique is faster, less labor-intensive and more reliable in comparison to other molecular typing methods (Bostock et al., 1993; Sikora et al., 1997) and thus was found to be more feasible method for this study.

Considering the high relevance of the above facts, the present study was designed to determine the genetic diversity among 177 Bt strains of Bangladesh and to reveal the correlation between the distribution pattern of cry genes to the diversity.

1 Results and Discussion
1.1 RAPD profile based genotyping
The specific typing of B. thuringiensis enables tracking of strains dispersed in the environment and assist in the discovery of new strains. The existing serotyping scheme, while having provided an invaluable basis for Bt classification for a long time, provides no information about the genetic relatedness of strains within groups and between groups and does not necessarily indicates the degree and spectrum of toxicity (Gaviria- Rivera and Priest, 2003). Contrarily, RAPD is a modified PCR method with a single arbitrary primer that recognizes differences in the prevalence and positions of annealing sites in the genome producing a spectrum of amplicons that are considered to reflect the genomic composition of the strain and may vary along the strains (Welsh and McClelland, 1990; Williams et al., 1990). The advantage of this method is that no prior knowledge of the genome under research is necessary (Bostock et al., 1993; Sikora et al., 1997) and it was found from several studies that the RAPD analysis could effectively distinguish between the Bt strains (Kumar et al., 2010). Hence, the analysis of genetic diversity among the indigenous Bt strains was done by RAPD-PCR method in this study.

A total of 177 Bt strains were employed for RAPD-PCR amplification with decamer OPA 03 (5'- AGCTCAGCCA -3') which was reported to produce 100% polymorphism (Kumar et al., 2010) and also observed to be efficient in few initial screenings in this study (Data not shown). Molecular weights of these bands were then estimated comparing with the DNA standard (100 bp DNA ladder, Bioneer, Korea) and the presence of 16 different bands (100, 150, 200, 225, 275, 300, 350, 400, 450, 500, 600, 700, 800, 900, 1000 and 1100 bp) were observed at varied numbers and combinations in the strains. By scrutinizing these bands, binary matrices for the strains were obtained where the presence was scored with 1 and absence with 0. No strain was observed with all 16 bands and maximum 11 bands were present for certain strains. Polymorphism based on these 16 individual bands was calculated and 100% polymorphism was not observed in any strain. Maximum 68.8% polymorphism was seen in 0.4% of the test strains whereas 25% polymorphism was most prevalent followed by 31.3% (15.3% of test strains) and 18.8% (13.1% of test strains) (Table 1).

As binary matrix was prepared for each strain based on the 16 polymorphic DNA bands, 256 (162) numbers of different banding patterns are possible. Therefore, 15 major RAPD pattern i.e. 15 genotyps were presumed from all these binary matrices with proximities to the major patterns which was derived from the dendrogram based on the heights of the clades (Fig 1- sup). Dendrogram (Fig 1- sup) was constructed by UPGMA clustering method with the distance matrix and similarity matrix among the strains prepared by dice coefficient comparison method from their binary matrices i.e. the numerical RAPD profiles (Table 1- sup). The heights of the clades in scale bar, an indication of distance among the strains was standardized in this case. A middle height at 0.2 was presumed as the threshold point from the scale bar where the start point extended up to 0.45 to distinguish the major clusters as separate genotype throughout the whole study. Thus the height of any clade (cluster) exceeding 0.2 (Fig 1-sup) was considered as separate genotype. RAPD pattern of Bt strains representing these 15 genotypes are as shown in Figure 1. This enabled quantitative comparison of genetic diversities among different sets of strains e.g. biotypes or locations upon standardization.

The Bt strains were found to be divided into two major clusters, A and B in the dendrogram (Fig 1- sup). Cluster B was smaller comprising of only 14 strains and A was larger with 163 strains, hence subdivided further into subclusters A1 and A2. Sub-cluster A2 was large enough and found to be further branched into clusters with significant number of strains denoted as A2a and A2b.

Genotype 2, 4 and 10 were simplicifolious (single leafed), genotype 14 was bifolious (two leaved), genotype 8 and 15 were trifolious (three leaved) and rest others were polyfolious (more than three leaved). The genotypes were observed to contain the strains in a mingling manner with respect to their biotypes except genotype 12 which contained 80% of strains from biotype kurstaki. And for the locations, strains from the same location appeared to be closely related even though their biochemical characteristics differed (Fig 1- sup).
 

The prevalence of Bt strains in different genotype was also calculated and genotype 9 and 11 were found to be the largest, each containing more than 25% of the strains (Fig 2).
 

1.2 Comparison of diversity between biotypes and locations
Based on the threshold height or distance of the clades in the scale bar, the diversity indices (DI) were calculated as the ratio between the number of major clusters and the number of strains, within biotypes and locations. In case of biotypes, genetic diversity was maximum in Bt israelensis followed by sotto, eleven and minimum was in biotype 13, ten and nine (Fig 3A). In case of locations, maximum diversity was observed among the strains of Narshingdi and the minimum was for Munshigonj (Fig 3B).
 

The average diversity index for locations (0.27±0.098) was higher than that for biotypes (0.23±0.046) which indicates that the genetic diversity among the strains of a certain location is not resulted from the influence of abiotic factors only such as UV, salinity, trace elements, pH, organic maters etc rather a phenotypical pattern was found to be maintained as the DI among the strains with similar biochemical properties was found to be lower across different locations.

1.3 Prevalence of cry genes
Detection of major cry genes (viz. cry1, cry2, cry3, cry4, cry8, cry9, cry10 and cry11) in the indigenous Bt strains was reported earlier (Shishir et al., 2014). Besides the prevalence of primary subgroups (Fig 4A), tertiary subgroups of cry1 gene (cry1Aa, cry1Ac, cry1Ba, cry1Ca belonging to cry1 as it was the most prevalent) were also detected (Fig 2A- sup, 2B- sup, 2C- sup and 2D- sup) and their prevalence was determined where cry1Aa and cry1Ca were found to be the most prevalent (Fig 4B).
 

1.4 Pattern of cry genes distribution within RAPD-genotyes
As the genotypes and cry gene profiles of the strains were thus retrieved, it was analyzed whether the distribution of cry genes was random or genotype oriented. So, the distribution of different cry genes in different genotypes was analyzed and from the graphical presentation (Fig 5) it was revealed that cry genes were present in all genotypes except genotype 10. The abundance of these cry genes were maximum in genotypes 1, 6, 9, 11 and 12. Though genotypes 9 and 11 were found to be the largest containing more than 25% of the strains, only genotype 9 of them was significant with different cry genes besides genotypes 1 and 6 as compared to the number of strains.
 

Comparing the ratio between the number of cry genes and strains in the genotypes, genotype 6 (2.167) was found to be most significant followed by genotype 1 (1.285), genotype 9 (0.29), genotype 11 (0.18) and genotype 3 (0.14). On the other hand, maximum 6 types of cry genes were present in genotypes 1, 6, 9 and 11. Thus, it was clear from this analysis that though the cry genes were observed in varied frequencies in most of the genotypes, cry genes were found to be most abundant in terms of number and type in genotype 6, 1 and 9. Again, the presence of same cry gene in different genotypes increased the chances that the degree and spectrum of toxicity might be variable i.e. genes except cry4, cry8 and cry10 were found to be present in multiple genotypes.
 
1.5 Comparison between different similarity parameters
Another comparison was performed with 20 Bt strains (indigenous- 19, reference- 1) in terms of their 16S rRNA gene sequence based phylogeny (Shishir et al., 2014), Biotype, RAPD based genotype and number of available cry genes (Fig 6) which revealed that phylogenetically close strains were similar in biochemical properties. Phylogenetically close 12 strains as in sub-cluster A1 were observed to have similar biochemical properties since from same biotype kurstaki except strain DSf3 (non-hemolytic), strain CiSa5 (biotype ten) and KSa2 (dendrolimus). Again in sub-cluster A2, 3 strains out of 5 were non- hemolytic and 2 out of 3 strains in cluster B were from biotype kurstaki. Though the biochemical properties of most of them conformed to the phylogenetic relatedness, their RAPD genotypes were variable e. g. 5 genotypes (1, 5, 6, 11 and 12) were visible among the strains of sub-cluster A1 which were from the biotype kurstaki. This genetic diversity might be due to the presence of many different plasmids in each strain and high frequency of DNA rearrangements in variable regions by conjugation transfer mechanism and the transposon-like inverted repeats flanking the endotoxin genes. Plasmid DNA exchange in nature is well documented in B. thuringiensis strains and has been implicated as the source of the remarkable diversity of cry genes (Carlson and Kolstø, 1993). Other comparisons of soil isolates of B. cereus and B. thuringiensis strains, selected with no regard for insect toxicity, have demonstrated extensive chromosomal DNA exchange with no apparent clonal population structure (Hu et al., 2004). On the other hand, correlation persisted for the highly conserved phenotypes like biochemical properties and genotypes such as 16S rRNA etc. as these are regulated by the in house conserved genes.
 

The number of available cry genes among these strains was also variable. It can, therefore, be said thatthe report of conformity between phylogenetic and phenotypic i.e. biotype or serotype (biotype in this case) relatedness (Shishir et al., 2014) was also evidenced in this study though RAPD- genotyping and cry gene profile did not follow the pattern.

2 Materials and methods
2.1 Bacterial strains and growth conditions
Indigenous Bt strains (n=177) preserved at Bt resource Centre in the Department of Microbiology, University of Dhaka and the reference Bt kurstaki HD-73, Bt sotto T84A1 and Bt japonensis Buibui strains collected from Bt stock collection of Okayama University, Japan were used in this study. LB agar (per litre: tryptone 10 g, yeast extract 5 g, NaCl 10 g, agar 15 g) and LB broth were used for culture maintenance, propagation and subculture throughout the study.
 
2.2 Total DNA extraction
Total DNA was extracted from the indigenous Bt isolates streaked on LB agar medium (Bravo et al., 1998). After 12 hours of incubation at 30°C, a single colony, transferred into 100 µl of sterile de-ionized water in a microfuge tube, was vortexed and kept at -70°C for 30 min. It was then incubated in boiling water for 10 min to lyse the cells and briefly centrifuged for 20 s at 12,000×g (Eppendorf centrifuge, 5415D). The upper aqueous phase transferred into sterile microfuge tubes was used as template and preserved at -30°C for further use. 50-100 ng of DNA from this suspension was used as template in RAPD- PCR analysis.

2.3 RAPD-PCR analysis
RAPD-PCR was performed using the primer (OPA 03: 5'- AGCTCAGCCA- 3') with minimum 60% G+C content and devoid of any internal repeat as maximum polymorphism was reported for this decamer (Kumar et al., 2010). PCR was carried out within a reaction volume of 25 µl [1× PCR Master mix (Promega, USA), 2.0 μM of primer, 50-100 ng of template DNA] in a thermal cycler (MJ mini, BioRad, USA) by 35 cycles (95°C for 1 min, 40°C for 1 min, 72°C for 1 min) with an initial denaturation step at 95°C for 4 min and a final extension step at 72°C for 15 min (Kumar et al., 2010). PCR products (15 μl) were then analyzed in 1.5% (w/v) Agarose (Promega, USA) gel by horizontal electrophoresis at 60V for 1h in 1× TBE [89 mM Tris (pH 7.6), 89 mM boric acid, 2 mM EDTA] buffer and gel images were captured after visualization against UV trans-illumination in a gel documentation system (Alpha imager mini, USA) following staining in Ethidium Bromide (EtBr) (Sigma, USA) solution (0.5 μg/ ml) and destaining in distilled water. Molecular weight of the DNA bands in those gels was then determined by using Alphaview SA software (version 3.4.0.0).

2.4 Data analysis and dendrogram construction
Binary matrix was prepared for each strain from the gel images based on the presence or absence as scored 1 or 0 respectively for the amplicons bands. Based on the binary matrices, similarity and distance matrices were calculated following dice coefficient method. These data were used in cluster analysis by UPGMA method to construct the dendrograms (http://insilico.ehu.es/dice_upgma/). For maximum accuracy of comparison, all isolates were processed with the same batch of PCR master mix.

2.5 Genotyping and estimation of diversity index
Throughout the whole study, threshold level was chosen at 0.2 in the scale bar. Each cluster was considered a separate genotype if distances among the strains in that cluster were less than 0.2 in scale bar. Thus the genotypes were identified among the tested strains as a whole and also in terms of biotype and location. Again, the ratio between the number of clusters and isolates for a set of strains was considered as their diversity index. Based on this criterion, the diversity index for biotypes and locations were estimated and compared.

2.6 Detection of subgroups of cry1 genes
DNA templates (50-100 ng) from the Bt strains (n= 177) were mixed with PCR reaction mixture containing 0.5 µM of each primer, 1× PCR Master mix (Promega, USA) in 25 μl reaction volume and amplification was performed in a DNA thermal cycler to detect cry genes belonging to the cry1 family (major group) such as cry1Aa, cry1Ac, cry1Ba, cry1Ca. For all primer sets, PCR was carried out with an initial single denaturation step at 95°C for 2 min and 30 amplification cycles including denaturation at 95°C for 45 s, annealing at 53°C for 45 s and extension at 72°C for 60 s. Finally an extra extension step was applied at 72°C for 10 min. PCR products (10 μl) were then electrophoresed in 1.5% agarose (Promega, USA) gel prepared and submerged in 1×TBE buffer at 60V for 60 min. Gel was visualized in a gel documentation system following staining and de-staining.

2.7 Determination of distribution of cry genes in the genotypes
Detection of cry genes from major groups such as cry1, cry2, cry3, cry4, cry8, cry9 and cry10 was reported previously (Shishir et al., 2014). Certain other subgroups of cry genes belonging to the cry1 family (major group) such as cry1Aa, cry1Ac, cry1Ba, cry1Ca were also investigated. Combining the presence of the above mentioned genes, cry gene profile was obtained for each strains. Again, each strain for its RAPD profile belongs to a certain genotype. Thus, the number of cry genes in each genotype and their distribution was determined.
 
3 Conclusion
This study will facilitate the research on Bacillus thuringiensis in Bangladesh by providing valuable information to find out diverse Bt strains those which can be used in biotechnological purposes. The results also suggest that the degree and spectrum of the toxicity of indigenous Bt strains could be diverse to be used as efficient weapons to fight the resistance problems with pests.

Acknowledgements
This work was supported by a Grant-in-Aid from the USDA as a project entitled "Production of Bacillus thuringiensis biopesticides by biotechnological approach for the control of vegetable pests in Bangladesh". We thank Okayama University, Japan for providing reference strain.

References
Adang, M. J., Crickmore, N. and Jurat-Fuentes, J. L. 2014, Diversity of Bacillus thuringiensis Crystal Toxins and Mechanism of Action, p. 39-87. In: T. S. Dhadialla and S. S. Gill (ed.). Advances in Insect Physiology. Oxford, Academic Press, London, UK
http://dx.doi.org/10.1016/b978-0-12-800197-4.00002-6

Bostock, A., Khattak, M. N., Matthews, R. and Burnie, J. 1993, Comparison of PCR fingerprinting, by random amplification of polymorphic DNA, with other molecular typing methods for Candida albicans, J. Gen. Microbiol. 139(9), 2179-2184
http://dx.doi.org/10.1099/00221287-139-9-2179 PMid:7902410

Bourque, S. N., Valero, J. R., Lavoie, M. C. and Levesque, R. C. 1995, Comparative analysis of the 16S to 23S ribosomal intergenic spacer sequences of Bacillus thuringiensis strains and subspecies and of closely related species, Appl. Environ. Microb. 61(4), 1623-1626.

Brar, S. K., Verma, M., Tyagi, R. D., Valero, J. R. and Surampalli, R. Y. 2009, Entomotoxicity, protease and chitinase activity of Bacillus thuringiensis fermented wastewater sludge with a high solids content, Bioresource technol. 100(19), 4317-4325.
http://dx.doi.org/10.1016/j.biortech.2007.09.093 PMid:19447031

Bravo, A., Sarabia, S., Lopez, L., Ontiveros, H., Abarca, C., Ortiz, A., Ortiz, M., Lina, L., Villalobos, F. J., Pena, G., Nunez-Valdez, M. E., Soberon, M. and Quintero, R. 1998, Characterization of cry genes in a Mexican Bacillus thuringiensis strain collection, Appl. Environ. Microb. 64(12), 4965-4972.

Brousseau, R., Saint-Onge, A., Prefontaine, G., Masson, L. and Cabana, J. 1993, Arbitrary primer polymerase chain reaction, a powerful method to identify Bacillus thuringiensis serovars and strains, Appl. Environ. Microb. 59(1), 114-119.

Bruce, M. J., Gatsi, R., Crickmore, N. and Sayyed, A. H. 2007, Mechanisms of resistance to Bacillus thuringiensis in the Diamondback Moth, Biopesticide International 3(1), 1-12.

Carlson, C. R. and Kolst?, A. -B. 1993, A complete physical map of Bacillus thuringiensis chromosome, J. Bacteriol. 175, 1053-1060.

Crickmore, N., Baum, J., Bravo, A., Lereclus, D., Narva, K., Sampson, K., Schnepf, E., Sun, M. and Zeigler, D. R. 2014, Bacillus thuringiensis toxin nomenclature. http://www.btnomenclature.info/ 

 Gominet, M., Slamti, L., Gilois, N., Rose, M. and Lereclus, D. 2001, Oligopeptide permease is required for expression of the Bacillus thuringiensis plcR regulon and for virulence, Mol. microbial. 40(4), 963-975.

Gaviria- Rivera, A. M., and Priest, F. G. (2003) Pulsed field gel electrophoresis of chromosomal DNA reveals a clonal population structure to Bacillus thuringiensis that relates in general to crystal protein gene content, FEMS Microbiol Lett 223, 61-66.
http://dx.doi.org/10.1016/S0378-1097(03)00347-1

Hajaij-Ellouze, M., Fedhila, S., Lereclus, D. and Nielsen-LeRoux, C. 2006, The enhancin-like metalloprotease from the Bacillus cereus group is regulated by the pleiotropic transcriptional activator PlcR but is not essential for larvicidal activity, FEMS microbiol. Lett. 260(1), 9-16.
http://dx.doi.org/10.1111/j.1574-6968.2006.00289.x PMid:16790012

Hu, X., Hansen, B. M., Eilenberg, J., Hendriksen, N. B., Smidt, L., Yuan, Z., and Jensen, G. B. 2004, Conjugative transfer, stability and expression of a plasmid encoding a cry1Ac gene in Bacillus cereus group strains, FEMS Microbiol. Lett. 231, 45-52.
http://dx.doi.org/10.1016/S0378-1097(03)00925-X

Infante, I., Morel, M. A., Ubalde, M. C., Martínez-Rosales, C., Belvisi, S. and Castro-Sowinski, S. 2010, Wool-degrading Bacillus isolates: extracellular protease production for microbial processing of fabrics, World. J. Microb. Biot. 26(6), 1047-1052.
http://dx.doi.org/10.1007/s11274-009-0268-z

Jiang, H., Dong, H., Zhang, G., Yu, B., Chapman, L. R. and Fields, M. W. 2006, Microbial diversity in water and sediment of Lake Chaka, an athalassohaline lake in northwestern China, Appl. Environ. Microb. 72(6), 3832-3845.
http://dx.doi.org/10.1128/AEM.02869-05PMid:16751487,PMCid:PMC1489620

Kaur, S. 2006, Molecular approaches for identification and construction of novel insecticidal genes for crop protection, World. J. Microb. Biot. 22(3), 233-253.
http://dx.doi.org/10.1007/s11274-005-9027-y

Kumar, D., Chaudhary, K. and Boora, K. S. 2010, Characterization of native Bacillus thuringiensis strains by PCR-RAPD based fingerprinting, Indian J. Microbiol. 50(1), 27-32.
http://dx.doi.org/10.1007/s12088-009-0011-3,PMid:23100804,PMCid:PMC3450284

Martin, P. A., Gundersen-Rindal, D. E. and Blackburn, M. B. 2010, Distribution of phenotypes among Bacillus thuringiensis strains, Syst. Appl. Microbiol. 33(4), 204-208.
http://dx.doi.org/10.1016/j.syapm.2010.04.002 PMid:20447792

Miteva, V., Abadjieva, A. and Grigorova, R. 1991, Differentiation among strains and serotypes of Bacillus thuringiensis by M13 fingerprinting, J. Gen. Microbiol. 137, 593-600.
http://dx.doi.org/10.1099/00221287-137-3-593

Nisnevitch, M., Sigawi, S., Cahan, R. and Nitzan, Y. 2010, Isolation, characterization and biological role of camelysin from Bacillus thuringiensis subsp. israelensis, Curr. Microbiol. 61(3), 176-183.
http://dx.doi.org/10.1007/s00284-010-9593-6 PMid:20127334

Palvannan, T. and Boopathy, R. 2005, Phosphatidylinositol-Specific Phospholipase C Production from Bacillus thuringiensis Serovar. kurstaki using Potato-Based Media, World. J. Microb. Biot. 21(6-7), 1153-1155.
http://dx.doi.org/10.1007/s11274-005-0299-z

Shishir, A., Roy, A., Islam, N., Rahman, A., Khan, S. N. and Hoq, M. M. 2014, Abundance and diversity of Bacillus thuringiensis in Bangladesh and their cry genes profile, Front. Environ. Sci. 2, 20.
http://dx.doi.org/10.3389/fenvs.2014.00020

Shishir, M. A., Akter, A., Bodiuzzaman, M., Hossain, M. A., Alam, M. M., Khan, S. A., Khan, S. N. and Hoq, M. M. 2015, Novel toxicity of Bacillus thuringiensis strains against melon fruit fly, Bactrocera cucurbitae (Diptera: Tephritidae), Biocontrol Sci. 20(2), 115-123.
http://dx.doi.org/10.4265/bio.20.115 PMid:26133509

Sikora, S., Redzepovic, S., Pejic, I. and Kozumplik, V. 1997, Genetic diversity of Bradyrhizobium japonicum field population revealed by RAPD fingerprinting, J. Appl. Microbiol. 82(4), 527-531.
http://dx.doi.org/10.1046/j.1365-2672.1997.00140.x

te Giffel, M. C., Beumer, R. R., Klijn, N., Wagendorp, A. and Rombouts, F. M. 1997, Discrimination between Bacillus cereus and Bacillus
thuringiensis using specific DNA probes based on variable regions of 16S rRNA, FEMS Microbiol. Lett. 146(1), 47-51
http://dx.doi.org/10.1111/j.1574-6968.1997.tb10169.x
http://dx.doi.org/10.1016/S0378-1097(96)00439-9

Welsh, J. and McClelland, M. 1990, Fingerprinting genomes using PCR with arbitrary primers, Nucleic Acids Res. 18(24), 7213-7218
http://dx.doi.org/10.1093/nar/18.24.7213,PMid:2259619,PMCid:PMC332855

Williams, J. G. K., Kubelik, A. R., Livak, K. J., Rafalski, J. A., Tingey, S. V. 1990, DNA polymorphisms amplified by arbitrary primers are useful as genetic markers. Nucleic Acids Res. 18, 6531-6535
http://dx.doi.org/10.1093/nar/18.22.6531,PMid:1979162,PMCid:PMC332606