Genetic networking of the Bemisia tabaci cryptic species complex reveals pattern of biological invasions

Abstract

Background: A challenge within the context of cryptic species is the delimitation of individual species within the complex. Statistical parsimony network analytics offers the opportunity to explore limits in situations where there are insufficient species-specific morphological characters to separate taxa. The results also enable us to explore the spread in taxa that have invaded globally.

Methodology/principal findings: Using a 657 bp portion of mitochondrial cytochrome oxidase 1 from 352 unique haplotypes belonging to the Bemisia tabaci cryptic species complex, the analysis revealed 28 networks plus 7 unconnected individual haplotypes. Of the networks, 24 corresponded to the putative species identified using the rule set devised by Dinsdale et al. (2010). Only two species proposed in Dinsdale et al. (2010) departed substantially from the structure suggested by the analysis. The analysis of the two invasive members of the complex, Mediterranean (MED) and Middle East - Asia Minor 1 (MEAM1), showed that in both cases only a small number of haplotypes represent the majority that have spread beyond the home range; one MEAM1 and three MED haplotypes account for >80% of the GenBank records. Israel is a possible source of the globally invasive MEAM1 whereas MED has two possible sources. The first is the eastern Mediterranean which has invaded only the USA, primarily Florida and to a lesser extent California. The second are western Mediterranean haplotypes that have spread to the USA, Asia and South America. The structure for MED supports two home range distributions, a Sub-Saharan range and a Mediterranean range. The MEAM1 network supports the Middle East - Asia Minor region.

Conclusion/significance: The network analyses show a high level of congruence with the species identified in a previous phylogenetic analysis. The analysis of the two globally invasive members of the complex support the view that global invasion often involve very small portions of the available genetic diversity.

Figure 1. Networks 1–11 as per Table 1 .

Network analysis based on statistical parsimony showing the genealogical relationships of the COI haplotypes in cryptic species of Bemisia tabaci. The Dinsdale putative species are indicated in parentheses. Networks encompassed by a box are those Dinsdale putative species that the analysis suggests have additional species level separation. The names codes in the ovals refer to the individual haplotypes (see Table S1). The small circles indicate the presence of missing intermediates while the connections are based on the set of plausible solutions with a 95% of parsimony probability. The number of sequences for each haplotype where n>1 is indicated next to the node; nodes without a number are n = 1.

Figure 2. Networks 12–29 and 35 as per Table 1 .

Network analysis based on statistical parsimony showing the genealogical relationships of the COI haplotypes in cryptic species of Bemisia tabaci. The Dinsdale putative species are indicated in parentheses. Networks encompassed by a box are those Dinsdale putative species that the analysis suggests have additional species level separation. The names codes in the ovals refer to the individual haplotypes (see Table S1). The small circles indicate the presence of missing intermediates while the connections are based on the set of plausible solutions with a 95% of parsimony probability. The number of sequences for each haplotype where n>1 is indicated next to the node; nodes without a number are n = 1. Networks 31–24 which related to the Mediterranean (networks 30, 31) and Middle East - Asia Minor 1 (networks 32–34) putative species are shown in Figs 3 and 4, respectively.

Figure 3. Network analysis based on statistical parsimony showing the genealogical relationships of the COI haplotypes representing the Mediterranean putative species.

The two networks are numbered 30 and 31. Network 30 represents 85 of the 86 haplotypes assigned to MED using the approach outlined in . The names codes in the ovals refer to the individual haplotypes (see Table S1). The small circles indicate the presence of missing intermediates while the connections are based on the set of plausible solutions with a 95% of parsimony probability. Each of the haplotypes (23 in total, see Table 4) for which there were at least two identical representatives in GenBank are indicated by an arrow linked to a box which details the countries where the haplotype was found; countries that are in bold and underlined are those which indicate an invasion. The ovals coloured in red indicate those haplotypes that together represent >80% of the GenBank records in the invaded ranges. The number of sequences for each haplotype where n>1 is indicated next to the node; nodes without a number are n = 1.

Figure 4. Network analysis based on statistical parsimony showing the relationships of the mtCOI haplotypes representing the Middle East – Asia Minor 1 putative species.

The three networks are numbered 32, 33 and 34. Network 32 represents 79 of the 80 haplotypes assigned to MEAM1 using the approach outlined in Dinsdale et al. (2010) . The name codes in the ovals refer to the individual haplotypes (see Table S1). The small circles indicate the presence of missing intermediates while the connections are based on the set of plausible solutions with a 95% of parsimony probability. Each of the haplotypes (11 in total, see Table 5) for which there were at least two identical representatives in GenBank are indicated by an arrow linked to a box which details the countries where the haplotype was found; countries that are in bold and underlined are those which indicate an invasion. The oval coloured in red indicates the haplotype that represents >80% of the GenBank records in the invaded ranges. The number of sequences for each haplotype where n>1 is indicated next to the node; nodes without a number are n = 1.