Limits...
Herpes simplex virus-2 genital tract shedding is not predictable over months or years in infected persons.

Dhankani V, Kutz JN, Schiffer JT - PLoS Comput. Biol. (2014)

Bottom Line: Empirical and model-derived, time-series data gathered over >30 days consists of multiple complex episodes that could not be reduced to a manageable number of descriptive features with PCA and SVD.However, single HSV-2 shedding episodes, even those with prolonged duration and complex morphologies consisting of multiple erratic peaks, were consistently described using a maximum of four dominant features.These findings suggest that HSV-2 shedding patterns within an individual are impossible to predict over weeks or months, and that even highly complex single HSV-2 episodes can only be partially predicted based on spatial distribution of immune cell density.

View Article: PubMed Central - PubMed

Affiliation: Department of Applied Mathematics, University of Washington, Seattle, Washington, United States of America.

ABSTRACT
Herpes simplex virus-2 (HSV-2) is a chronic reactivating infection that leads to recurrent shedding episodes in the genital tract. A minority of episodes are prolonged, and associated with development of painful ulcers. However, currently, available tools poorly predict viral trajectories and timing of reactivations in infected individuals. We employed principal components analysis (PCA) and singular value decomposition (SVD) to interpret HSV-2 genital tract shedding time series data, as well as simulation output from a stochastic spatial mathematical model. Empirical and model-derived, time-series data gathered over >30 days consists of multiple complex episodes that could not be reduced to a manageable number of descriptive features with PCA and SVD. However, single HSV-2 shedding episodes, even those with prolonged duration and complex morphologies consisting of multiple erratic peaks, were consistently described using a maximum of four dominant features. Modeled and clinical episodes had equivalent distributions of dominant features, implying similar dynamics in real and simulated episodes. We applied linear discriminant analysis (LDA) to simulation output and identified that local immune cell density at the viral reactivation site had a predictive effect on episode duration, though longer term shedding suggested chaotic dynamics and could not be predicted based on spatial patterns of immune cell density. These findings suggest that HSV-2 shedding patterns within an individual are impossible to predict over weeks or months, and that even highly complex single HSV-2 episodes can only be partially predicted based on spatial distribution of immune cell density.

Show MeSH

Related in: MedlinePlus

Low rank approximations for sample episodes from (a) Modeled data (10 episodes, 100 swabs/day) and (b) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight) using the four most dominant features obtained from SVD.For the 3 clinical datasets (different sampling rates) and 1 model simulation dataset, rank-4 approximations were >85% accurate.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4222642&req=5

pcbi-1003922-g004: Low rank approximations for sample episodes from (a) Modeled data (10 episodes, 100 swabs/day) and (b) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight) using the four most dominant features obtained from SVD.For the 3 clinical datasets (different sampling rates) and 1 model simulation dataset, rank-4 approximations were >85% accurate.

Mentions: We next switched focus to individual episodes of viral reactivation. Many episodes last fewer than 12 hours and are characterized by single expansion and decay phases and would therefore be easily classified with a minimal number of variables. Yet, the most clinically important episodes that are typically associated with lesions are defined by multiple erratic viral re-expansions and prolonged duration (0.5–3 weeks). In order to understand viral dynamics during these non-monotonically decaying episodes, we analyzed samples from four datasets: (a) Clinical data (23 episodes, swabs every 24 hours, [7], [13]), (b) Clinical data (5 episodes, swabs every 6 hours, [14]), (c) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight, [8]), and (d) Model simulation data (10 episodes, swabs every 15 minutes, [8]). Figure 3 shows representative curves of non-monotonically decaying episodes from the four datasets (left column: a, b, c and d), and plots of singular values obtained by applying SVD on the corresponding datasets (right column: e, f, g and h). The results were strikingly different than that for 60-day data that captured multiple diverse episodes. When SVD was applied to sets of individual non-monotonically decaying episodes, the most dominant variables accounted for 48.6, 60.8, 55.4 and 60.9 percent of the total variance of the 4 datasets respectively. To approximate these datasets to >90% accuracy required 11, 4, 5 and 6 variables respectively. Low-rank approximations with 100% accuracy was achieved with only 6 and 9 dominant variables when applied to the every 6 hour (Figure 3f) as well as the 10 swabs per day episodes (Figure 3g). Figure 4 demonstrates nearly perfect reproduction of single episodes generated with the mathematical model (Fig. 4a) and with 10 swabs per day (Fig. 4b) with only the 4 most dominant modes.


Herpes simplex virus-2 genital tract shedding is not predictable over months or years in infected persons.

Dhankani V, Kutz JN, Schiffer JT - PLoS Comput. Biol. (2014)

Low rank approximations for sample episodes from (a) Modeled data (10 episodes, 100 swabs/day) and (b) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight) using the four most dominant features obtained from SVD.For the 3 clinical datasets (different sampling rates) and 1 model simulation dataset, rank-4 approximations were >85% accurate.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4222642&req=5

pcbi-1003922-g004: Low rank approximations for sample episodes from (a) Modeled data (10 episodes, 100 swabs/day) and (b) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight) using the four most dominant features obtained from SVD.For the 3 clinical datasets (different sampling rates) and 1 model simulation dataset, rank-4 approximations were >85% accurate.
Mentions: We next switched focus to individual episodes of viral reactivation. Many episodes last fewer than 12 hours and are characterized by single expansion and decay phases and would therefore be easily classified with a minimal number of variables. Yet, the most clinically important episodes that are typically associated with lesions are defined by multiple erratic viral re-expansions and prolonged duration (0.5–3 weeks). In order to understand viral dynamics during these non-monotonically decaying episodes, we analyzed samples from four datasets: (a) Clinical data (23 episodes, swabs every 24 hours, [7], [13]), (b) Clinical data (5 episodes, swabs every 6 hours, [14]), (c) Clinical data (8 episodes, 10 swabs/day: every 2 hours during the day and 4 hours overnight, [8]), and (d) Model simulation data (10 episodes, swabs every 15 minutes, [8]). Figure 3 shows representative curves of non-monotonically decaying episodes from the four datasets (left column: a, b, c and d), and plots of singular values obtained by applying SVD on the corresponding datasets (right column: e, f, g and h). The results were strikingly different than that for 60-day data that captured multiple diverse episodes. When SVD was applied to sets of individual non-monotonically decaying episodes, the most dominant variables accounted for 48.6, 60.8, 55.4 and 60.9 percent of the total variance of the 4 datasets respectively. To approximate these datasets to >90% accuracy required 11, 4, 5 and 6 variables respectively. Low-rank approximations with 100% accuracy was achieved with only 6 and 9 dominant variables when applied to the every 6 hour (Figure 3f) as well as the 10 swabs per day episodes (Figure 3g). Figure 4 demonstrates nearly perfect reproduction of single episodes generated with the mathematical model (Fig. 4a) and with 10 swabs per day (Fig. 4b) with only the 4 most dominant modes.

Bottom Line: Empirical and model-derived, time-series data gathered over >30 days consists of multiple complex episodes that could not be reduced to a manageable number of descriptive features with PCA and SVD.However, single HSV-2 shedding episodes, even those with prolonged duration and complex morphologies consisting of multiple erratic peaks, were consistently described using a maximum of four dominant features.These findings suggest that HSV-2 shedding patterns within an individual are impossible to predict over weeks or months, and that even highly complex single HSV-2 episodes can only be partially predicted based on spatial distribution of immune cell density.

View Article: PubMed Central - PubMed

Affiliation: Department of Applied Mathematics, University of Washington, Seattle, Washington, United States of America.

ABSTRACT
Herpes simplex virus-2 (HSV-2) is a chronic reactivating infection that leads to recurrent shedding episodes in the genital tract. A minority of episodes are prolonged, and associated with development of painful ulcers. However, currently, available tools poorly predict viral trajectories and timing of reactivations in infected individuals. We employed principal components analysis (PCA) and singular value decomposition (SVD) to interpret HSV-2 genital tract shedding time series data, as well as simulation output from a stochastic spatial mathematical model. Empirical and model-derived, time-series data gathered over >30 days consists of multiple complex episodes that could not be reduced to a manageable number of descriptive features with PCA and SVD. However, single HSV-2 shedding episodes, even those with prolonged duration and complex morphologies consisting of multiple erratic peaks, were consistently described using a maximum of four dominant features. Modeled and clinical episodes had equivalent distributions of dominant features, implying similar dynamics in real and simulated episodes. We applied linear discriminant analysis (LDA) to simulation output and identified that local immune cell density at the viral reactivation site had a predictive effect on episode duration, though longer term shedding suggested chaotic dynamics and could not be predicted based on spatial patterns of immune cell density. These findings suggest that HSV-2 shedding patterns within an individual are impossible to predict over weeks or months, and that even highly complex single HSV-2 episodes can only be partially predicted based on spatial distribution of immune cell density.

Show MeSH
Related in: MedlinePlus