Fig 1.
Example of tracking a spermatozoon head motion based on a movie recorded at 10× magnification at 100fps for 0.5 s.
VSL (red), VAP (blue and green), VCL (black), ALH and BCF calculated for this track are stated (see text for the definitions). Coloured lines indicate the distance used in calculating the corresponding speed (see section on Tracking analysis for explanation of VAP). Inset: The oscillatory component of displacement ros plotted against time.
Fig 2.
Examples of theoretical f(q, τ) at q = 0.34μm−1 for typical parameters.
fm(q, τ) from Eq 7 for oscillatory motion (green): , f0 = 25 Hz, A0 = 3 μm; ballistic motion (red):
, σ = 70 μm s−1, f0 = 25 Hz, A0 = 0 μm; oscillatory and ballistic motion (orange):
, σ = 70 μm s−1, f0 = 25 Hz, A0 = 3 μm; diffusive motion (grey): fd(q, τ) for Dd = 0.3 μm2, and diffusive motion (black): fnm(q, τ) for Dnm = 0.03 μm2 s−1. Full f(q, τ) (blue): from all four (additive) contributions if swimming spermatozoa contribute 50% of the signal, debris 25% and non-motile spermatozoa 25%.
Fig 3.
Phase contrast micrographs (10×) of (left) an undiluted thawed sample with ∼ 80 × 106 cell/ml and (right) a diluted thawed sample with ∼ 20 × 106 cell/ml.
Scale bar = 100 μm.
Fig 4.
DDM results from a movie (2.5×, 100 fps, 480 p, 40 s) of a diluted sample.
(a) Measured g(q, τ) at 6 values of q, specified in the legend. Black lines are fits to the model given by Eq 2. The arrows indicate three processes associated with head oscillation (1), swimming (2), and diffusion of non-motile sperm cells (3). (b) Reconstructed f(q, τ) plotted against qτ and q2 τ (inset).
Fig 5.
Measured g(q, τ) from a movie (10×, 100 fps, 500 p, 200 s) of an in-active sample at q = 2.22μm−1 with a cell density of ≈ 20 × 16cells/ml.
Line is a fit using two exponential functions returning two separate diffusion coefficient D1 and D2. Right inset: fitted parameters D1 and D2 for a range of q. Left inset: Fitted measurement of the fractional contribution, γ, of the debris to the total signal.
Fig 6.
Fitted parameters over the range q = 0-0.4 μm−1.
(a) Mean and (b) width σ of the fitted Schultz swimming speed distribution, (c) proportion of motile cells α, (d) head amplitude A0, (e) head frequency f0 and (f) diffusion coefficient of the non-motile spermatozoa Dnm. (○) Fresh ejaculate diluted to ∼ 20 × 106 cells/ml and recorded using (2.5×, 100 fps, 480 p, 40 s). These parameters correspond to the data in Fig 4. (□) thawed semen undiluted (∼ 80 × 106 cells/ml) and recorded using (2×, 300 fps, 300 p, 40 s). (△) thawed semen diluted 4× to ∼ 20 × 106 cells/ml recorded using (2×, 1000 fps, 300 p, 8 s).
Fig 7.
A comparison of tracking and DDM methods for a sample maintained at 37°C for 15 min and 50 min after thawing.
(a,b) swimming speed distributions. Histogram of VAP (black) and VSL (grey) were calculated from 10 movies (10×, 100 fps, 500 p, 0.5 s) with ≈ 50 swimming tracks per movie. A lower magnification movie (2×, 300 fps, 500 p, 10 s) was recorded immediately afterwards and analysed with DDM to give (red vertical bar) and σ from which P(v) was reconstructed (red dashed line). Histograms are normalised such that the peak is at 1. (c,d) Head oscillation amplitude A0 and frequency f0 measured with DDM and tracking for consecutive movies of the same sample. Note that the quoted error for tracking is the standard deviation for measurements from 50 tracks, while that for DDM is a standard deviation of the mean for averaging over the values measured over a range of q values.
Table 1.
Motile fraction: DDM vs tracking.
Fig 8.
plotted against
for DDM measurements of thawed straws: BB (□), CH (△) and HO semen (○) were pipetted undiluted into pre-warmed sample chambers (≈ 80 × 106cells/ml) and movies (2×, 300 fps, 300 p, 40 s) were recorded immediately after filling every 2min subsequently. (Only the first 30 min of CH are included, before the data becomes noisy.) Data from Figs 6 and 7 and Table 1 plotted in (*). Grey area defines the range of linear fit through the origin for all three datasets and
.