A puzzle with many pieces: the genetic structure and diversity of Phaeocystis antarctica Karsten (Prymnesiophyta)

Figures & data

Fig. 1. Sampling sites of P. antarctica (white dots) and current systems in the Southern Ocean (redrawn from Rintoul et al., 2001). White arrows indicate the Antarctic Coastal Current (ACoC), acting as counter-current of the ACC.

Table 1. Number of alleles (N_a) and effective number of alleles (E_ffN_a), observed heterozygosity (H_o), expected heterozygosity (H_e, genetic diversity), inbreeding coefficient (G_is) and percentage of missing data per locus.

Locus	Na	EffNa	Ho	He	Gis	Missing [%]
Locus 1	33	3.4	0.63*	0.84	0.22	7.0
Locus 2	59	5.1	0.78*	0.97	0.14	12.8
Locus 3	39	4.3	0.71*	0.93	0.26	24.4
Locus 4	26	3.0	0.52*	0.78	0.32	2.3
Locus 5	41	3.1	0.70	0.91	0.09	18.6
Locus 6	51	4.3	0.65*	0.94	0.25	12.8
Locus 7	52	4.7	0.62*	0.98	0.39	31.4

* indicates significant deviations from HWP based on 999 random permutations.

Fig. 2. Number of unique genotypes when considering from one and up to seven microsatellites of the data set. Whereas all isolates shared genotypes for at least two loci, no isolate shared more than five of the analysed seven microsatellites indicating the great variation even within isolates from a single bucket of seawater.

Table 2. Results of partial Mantel tests.

Genetic distance measure	Gen × Geo [Time]	Gen × Time [Geo]
Smouse & Peakall	r = 0.1218, P = 0.007	r = 0.036, P = 0.0282
Chord	r = 0.1431, P < 0.001	r = 0.1156, P = 0.013
Manhattan	r = 0.142, P < 0.001	r = 0.0698, P = 0.082
Inverted Kinship	r = 0.1784, P < 0.001	r = 0.0712, P < 0.001
Euclidean	r = 0.1181, P = 0.009	r = 0.0262, P = 0.334

The first analysis tested for a correlation between the genetic (Gen) and geographic distance (Geo) controlling for time (Time). The second analysis tested for a correlation between genetic distance and time controlling for sampling location. The regression coefficient r and the significance (assessed via 10 000 randomizations) are listed. For partial Mantel tests data from 83 isolates with precise sampling station and time points (day/month/years) were used (see Supplementary Table 1).

Fig. 3. Assignment of individuals (coloured bars) to populations using full data set excluding two for which no sampling time information was available (86 isolates, see Supplementary Table 1) and the program STRUCTURE with population ‘time’ prior (npops = 11). Run parameters: 10 independent runs, burn-in 50 000 generations, MCMC 100 000 generations. In the diagrams below, the likelihood of the dataset for a given K (right) and Delta K, with the most likely number of K for the data set following Evanno et al. (2005) is shown. Both Delta K and the average likelihood for a given K were highest for K = 6 from the 10 independent runs.

Fig. 4. Assignment of individuals (coloured bars) to populations using full data set (with the exception of those where no sampling date was given, i.e. 85 isolates) and the program STRUCTURE without prior on population origin/time. Run parameters: 10 independent runs, burn-in 50 000 generations, MCMC 100 000 generations. Following the correction method of Evanno et al. (2005), the highest likelihood for a given K was for K = 4. Colours refer to the proportions of cluster for each genotype found.

Rintoul, S., Hughes, C. & Olbers, D. (2001). The Antarctic Circumpolar Current System. In Ocean Circulation and Climate (Siedler, G., Church, J. & Gould, J., editors), 271–302. Academic Press, New York.

 Google Scholar

Evanno, G., Regnaut, S. & Goudet, J. (2005). Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Molecular Ecology, 14: 2611–2620.

 PubMed Web of Science ®Google Scholar