Limits...
Winding up the molecular clock in the genus Carabus (Coleoptera: Carabidae): assessment of methodological decisions on rate and node age estimation.

Andújar C, Serrano J, Gómez-Zurita J - BMC Evol. Biol. (2012)

Bottom Line: Alternative choices of clock model, partitioning scheme, treatment of ambiguous characters, and outgroup inclusion resulted in rate increments ranging from 28% (HUWE1) to 1000% (LSU-B and ITS2) and increments in the TMRCA of Carabus ranging from 8.4% (cox1-A) to 540% (ITS2).The combination of several genes is proposed as the best strategy to minimise both the idiosyncratic behaviors of individual markers and the effect of analytical aspects in rate and age estimations.Our results highlight the importance of estimating rates of molecular evolution for each specific dataset, selecting for optimal clock and partitioning models as well as other methodological issues potentially affecting rate estimation.

View Article: PubMed Central - HTML - PubMed

Affiliation: Departamento de Zoología y Antropología Física, Facultad de Veterinaria, Universidad de Murcia, 30071 Murcia, Spain. candujar@um.es

ABSTRACT

Background: Rates of molecular evolution are known to vary across taxa and among genes, and this requires rate calibration for each specific dataset based on external information. Calibration is sensitive to evolutionary model parameters, partitioning schemes and clock model. However, the way in which these and other analytical aspects affect both the rates and the resulting clade ages from calibrated phylogenies are not yet well understood. To investigate these aspects we have conducted calibration analyses for the genus Carabus (Coleoptera, Carabidae) on five mitochondrial and four nuclear DNA fragments with 7888 nt total length, testing different clock models and partitioning schemes to select the most suitable using Bayes Factors comparisons.

Results: We used these data to investigate the effect of ambiguous character and outgroup inclusion on both the rates of molecular evolution and the TMRCA of Carabus. We found considerable variation in rates of molecular evolution depending on the fragment studied (ranging from 5.02% in cob to 0.26% divergence/My in LSU-A), but also on analytical conditions. Alternative choices of clock model, partitioning scheme, treatment of ambiguous characters, and outgroup inclusion resulted in rate increments ranging from 28% (HUWE1) to 1000% (LSU-B and ITS2) and increments in the TMRCA of Carabus ranging from 8.4% (cox1-A) to 540% (ITS2). Results support an origin of the genus Carabus during the Oligocene in the Eurasian continent followed by a Miocene differentiation that originated all main extant lineages.

Conclusions: The combination of several genes is proposed as the best strategy to minimise both the idiosyncratic behaviors of individual markers and the effect of analytical aspects in rate and age estimations. Our results highlight the importance of estimating rates of molecular evolution for each specific dataset, selecting for optimal clock and partitioning models as well as other methodological issues potentially affecting rate estimation.

Show MeSH

Related in: MedlinePlus

Ultrametric time-calibrated phylogenetic tree obtained with BEAST for nd5 data in Carabus. Nodes J1, J2, J3, J4b, M2, M3, C and F (labeled green) were used as calibration points (detailed information on node age priors as shown in Table 2). The HPD distribution of ages obtained for nodes A, B and C (labeled red) are used as priors for subsequent calibration analyses on the different individual and combined datasets.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC3368785&req=5

Figure 1: Ultrametric time-calibrated phylogenetic tree obtained with BEAST for nd5 data in Carabus. Nodes J1, J2, J3, J4b, M2, M3, C and F (labeled green) were used as calibration points (detailed information on node age priors as shown in Table 2). The HPD distribution of ages obtained for nodes A, B and C (labeled red) are used as priors for subsequent calibration analyses on the different individual and combined datasets.

Mentions: Calibration tests were standardized across the phylogenies estimated under different conditions, by using a fixed set of nodes constrained to absolute age intervals obtained from an independent analysis of an expanded nd5 dataset. This dataset included 51 Carabus and 7 outgroup nd5 sequences (Additional file 1: Table S2) and was analyzed in BEAST 1.5.4 [44] constraining eight calibrations points based on geologic events and the availability of a fossil Carabus from the end of the Miocene (Figure 1, Table 2). These data were investigated under different codon partition schemes (1P: no partitioning; 2P: two partitions, considering first and second codon positions together; 3P: each codon position as a different partition) and clock models, strict (SC) and uncorrelated lognormal (ULN) clocks. The optimal conditions were selected based on Bayes factors (BF) comparisons using marginal likelihood values as calculated in Tracer 1.5. Positive evidence based on BF was interpreted as requiring at least a ten units increase in marginal likelihood per additional free parameter before accepting a more complex model [45,46]. We assumed one extra parameter in ULN analyses compared to the SC assumption [20], and ten extra parameters per additional partition under a GTR + G + I model.


Winding up the molecular clock in the genus Carabus (Coleoptera: Carabidae): assessment of methodological decisions on rate and node age estimation.

Andújar C, Serrano J, Gómez-Zurita J - BMC Evol. Biol. (2012)

Ultrametric time-calibrated phylogenetic tree obtained with BEAST for nd5 data in Carabus. Nodes J1, J2, J3, J4b, M2, M3, C and F (labeled green) were used as calibration points (detailed information on node age priors as shown in Table 2). The HPD distribution of ages obtained for nodes A, B and C (labeled red) are used as priors for subsequent calibration analyses on the different individual and combined datasets.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC3368785&req=5

Figure 1: Ultrametric time-calibrated phylogenetic tree obtained with BEAST for nd5 data in Carabus. Nodes J1, J2, J3, J4b, M2, M3, C and F (labeled green) were used as calibration points (detailed information on node age priors as shown in Table 2). The HPD distribution of ages obtained for nodes A, B and C (labeled red) are used as priors for subsequent calibration analyses on the different individual and combined datasets.
Mentions: Calibration tests were standardized across the phylogenies estimated under different conditions, by using a fixed set of nodes constrained to absolute age intervals obtained from an independent analysis of an expanded nd5 dataset. This dataset included 51 Carabus and 7 outgroup nd5 sequences (Additional file 1: Table S2) and was analyzed in BEAST 1.5.4 [44] constraining eight calibrations points based on geologic events and the availability of a fossil Carabus from the end of the Miocene (Figure 1, Table 2). These data were investigated under different codon partition schemes (1P: no partitioning; 2P: two partitions, considering first and second codon positions together; 3P: each codon position as a different partition) and clock models, strict (SC) and uncorrelated lognormal (ULN) clocks. The optimal conditions were selected based on Bayes factors (BF) comparisons using marginal likelihood values as calculated in Tracer 1.5. Positive evidence based on BF was interpreted as requiring at least a ten units increase in marginal likelihood per additional free parameter before accepting a more complex model [45,46]. We assumed one extra parameter in ULN analyses compared to the SC assumption [20], and ten extra parameters per additional partition under a GTR + G + I model.

Bottom Line: Alternative choices of clock model, partitioning scheme, treatment of ambiguous characters, and outgroup inclusion resulted in rate increments ranging from 28% (HUWE1) to 1000% (LSU-B and ITS2) and increments in the TMRCA of Carabus ranging from 8.4% (cox1-A) to 540% (ITS2).The combination of several genes is proposed as the best strategy to minimise both the idiosyncratic behaviors of individual markers and the effect of analytical aspects in rate and age estimations.Our results highlight the importance of estimating rates of molecular evolution for each specific dataset, selecting for optimal clock and partitioning models as well as other methodological issues potentially affecting rate estimation.

View Article: PubMed Central - HTML - PubMed

Affiliation: Departamento de Zoología y Antropología Física, Facultad de Veterinaria, Universidad de Murcia, 30071 Murcia, Spain. candujar@um.es

ABSTRACT

Background: Rates of molecular evolution are known to vary across taxa and among genes, and this requires rate calibration for each specific dataset based on external information. Calibration is sensitive to evolutionary model parameters, partitioning schemes and clock model. However, the way in which these and other analytical aspects affect both the rates and the resulting clade ages from calibrated phylogenies are not yet well understood. To investigate these aspects we have conducted calibration analyses for the genus Carabus (Coleoptera, Carabidae) on five mitochondrial and four nuclear DNA fragments with 7888 nt total length, testing different clock models and partitioning schemes to select the most suitable using Bayes Factors comparisons.

Results: We used these data to investigate the effect of ambiguous character and outgroup inclusion on both the rates of molecular evolution and the TMRCA of Carabus. We found considerable variation in rates of molecular evolution depending on the fragment studied (ranging from 5.02% in cob to 0.26% divergence/My in LSU-A), but also on analytical conditions. Alternative choices of clock model, partitioning scheme, treatment of ambiguous characters, and outgroup inclusion resulted in rate increments ranging from 28% (HUWE1) to 1000% (LSU-B and ITS2) and increments in the TMRCA of Carabus ranging from 8.4% (cox1-A) to 540% (ITS2). Results support an origin of the genus Carabus during the Oligocene in the Eurasian continent followed by a Miocene differentiation that originated all main extant lineages.

Conclusions: The combination of several genes is proposed as the best strategy to minimise both the idiosyncratic behaviors of individual markers and the effect of analytical aspects in rate and age estimations. Our results highlight the importance of estimating rates of molecular evolution for each specific dataset, selecting for optimal clock and partitioning models as well as other methodological issues potentially affecting rate estimation.

Show MeSH
Related in: MedlinePlus