Computational protocol: Within-Host Speciation of Malaria Parasites

Similar protocols

Protocol publication

[…] Blood samples of wild birds were obtained at different European localities from southern Spain to Sweden, at migration stopover sites, and in European and African wintering areas. We screened 4513 wild birds in total, corresponding to 48 passerine species (47 European species plus the African hill babbler; ).We detected malaria infections by amplification of 479 bases of the parasite cytochrome b gene using DNA extracted from bird blood and highly efficient polymerase chain reaction (PCR) methods , . Different malaria lineages were distinguished by one or more nucleotide differences . Multiple infections revealed by mixed sequences were resolved by TA-cloning . In total, we scored 1927 infected birds, the average sample size being 40 infected individuals per species (median = 9.5, range 1–357; ). We found 143 distinct parasite lineages, each one being found in 1.9 species on average (range 1–22 species). In 13 blackcap cases (from distantly located sites or different years), we also amplified 220 bases of the nuclear DHFR-TS gene as described previously . All these sequences were retrieved from singly infected blackcaps, so that the association between parasite nuclear and mitochondrial DNA sequences could be unambiguously determined. We could not retrieve the sequence of this gene from all blackcap parasites because many occurred in mixed infections or were not amplified using our PCR . The DNA sequences used in this study have been deposited in GenBank ().We used PAUP to construct a maximum likelihood phylogenetic tree based on parasite cytochrome b sequences (), using a General Time Reversible model of nucleotide substitution with gamma parameter α = 0.623, and assumed proportion of invariable sites = 0.427. This was the best of 56 models according to the Akaike information criterion implemented in Modeltest . Support to internal branches was estimated by bootstrap analyses (1000 replicates) . We confirmed the tree by repeating the analysis using Bayesian methods as implemented in mrBayes 3.0 , under the same model of nucleotide substitution. This method produced the same tree topology, and similar or even stronger support for internal branches, as evaluated by posterior probabilities derived from trees sampled every 500 generations from a 10-million generations Markov Chain Monte Carlo series, with a burn-in time of 250000 generations that removed any trees generated before convergence had been reached. To construct the DHFR-TS tree, we used PAUP and a Kimura 3-parameters model with unequal base frequencies , . The exact probability of obtaining identical topologies for trees based on nuclear and mitochondrial genes was calculated by generating all possible trees with six leaves using COMPONENT 2.00a .We analysed the genetic structure of the blackcap parasite flock using an analysis of molecular variance (AMOVA) , comparing seven breeding populations covering most of the species' range, from southern Spain to Sweden (). The analysis used Kimura two-parameter distances under a gamma distribution with α = 0.12, as estimated from the data. The significance of the fixation index (ΦST) was tested by 5000 permutations of parasite haplotypes among populations .Aside from blackcaps, we extensively sampled 14 other species (n>40 birds and >25 scored infections), which were used to estimate intraspecific parasite richness (, ). To avoid sampling effects, curves of cumulative lineage richness (addition of new parasite lineages as new infected hosts are inspected) were constructed, and the number of parasite lineages found after scoring 25 infections was used as a standard estimate of parasite richness (R25). Average curves and R25 values (±S.E.) were derived from 1000 richness curves constructed by randomly changing the order in which individual hosts were screened. While the number of parasite lineages found in one species depended on the number of infections scored (r2 = 0.58, n = 48, P<0.0001), R25 was independent of sample size (r2 = 0.04, n = 15, P = 0.43). […]

Pipeline specifications

Software tools PAUP*, ModelTest-NG, MrBayes
Application Phylogenetics
Diseases Malaria