Evaluation of Dimensionality-Reduction Methods from Peptide Folding–Unfolding Simulations

Duan, Mojie; Fan, Jue; Li, Minghai; Han, Li; Huo, Shuanghong

doi:10.1021/ct400052y

Cited by 34 publications

(46 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…20 TRDG was also used as the "gold standard" to evaluate the calculation methods of configurational entropy 21 and the quality of dimensionality reduction. 22 However, the TRDG constructed from MD trajectories has its own limitations as shown in the section of Results, therefore, we propose to construct a weighted graph (G(V, E)) from the MD trajectories to depict the free energy landscape using discrete transition-path theory as a tool to analyze Markov state models. [23][24][25] Hereinafter, TRDG refers to the one constructed from MD trajectories.…”

Section: Introductionmentioning

confidence: 99%

Graph representation of protein free energy landscape

Duan

Fan

et al. 2013

The Journal of Chemical Physics

Self Cite

View full text Add to dashboard Cite

The thermodynamics and kinetics of protein folding and protein conformational changes are governed by the underlying free energy landscape. However, the multidimensional nature of the free energy landscape makes it difficult to describe. We propose to use a weighted-graph approach to depict the free energy landscape with the nodes on the graph representing the conformational states and the edge weights reflecting the free energy barriers between the states. Our graph is constructed from a molecular dynamics trajectory and does not involve projecting the multi-dimensional free energy landscape onto a low-dimensional space defined by a few order parameters. The calculation of free energy barriers was based on transition-path theory using the MSMBuilder2 package. We compare our graph with the widely used transition disconnectivity graph (TRDG) which is constructed from the same trajectory and show that our approach gives more accurate description of the free energy landscape than the TRDG approach even though the latter can be organized into a simple tree representation. The weighted-graph is a general approach and can be used on any complex system.

show abstract

Section: Introductionmentioning

confidence: 99%

Graph representation of protein free energy landscape

Duan

Fan

et al. 2013

The Journal of Chemical Physics

Self Cite

View full text Add to dashboard Cite

show abstract

“…Therefore, recent studies have been seeking to analyze MD simulations using high-performance computing to reduce the data dimensionality and capture the most important and effective dimensions from the simulation trajectories for such biological processes [1], [6], [17]. Thus, many dimensionality reduction techniques have been developed and employed to embed the 3 × N (N is the number of atoms in the molecule) dimensional conformational states in a low-dimensional latent space (2D or 3D) [18], [19]. The result embeddings are commonly projected into a 2D or 3D plot.…”

Section: A Backgroundmentioning

confidence: 99%

Visual Analytics for Deep Embeddings of Large Scale Molecular Dynamics Simulations

Chae

Bhowmik

et al. 2019

Preprint

View full text Add to dashboard Cite

Molecular Dynamics (MD) simulation have been emerging as an excellent candidate for understanding complex atomic and molecular scale mechanism of bio-molecules that control essential bio-physical phenomenon in a living organism. But this MD technique produces large-size and long-timescale data that are inherently high-dimensional and occupies many terabytes of data. Processing this immense amount of data in a meaningful way is becoming increasingly difficult. Therefore, specific dimensionality reduction algorithm using deep learning technique has been employed here to embed the high-dimensional data in a lower-dimension latent space that still preserves the inherent molecular characteristics i.e. retains biologically meaningful information. Subsequently, the results of the embedding models are visualized for model evaluation and analysis of the extracted underlying features. However, most of the existing visualizations for embeddings have limitations in evaluating the embedding models and understanding the complex simulation data. We propose an interactive visual analytics system for embeddings of MD simulations to not only evaluate and explain an embedding model but also analyze various characteristics of the simulations. Our system enables exploration and discovery of meaningful and semantic embedding results and supports the understanding and evaluation of results by the quantitatively described features of the MD simulations (even without specific labels).

show abstract

“…Surprisingly, Diffusion Maps, although a strong performer in the NPE method, shows a higher RV at every m than the other methods, and does not display an elbow effect. This is not necessarily an outcome of properties of the methods themselves, since Diffusion Maps was found to have a lower RV than Isomap and PCA at higher dimensions in a comparative analysis of dimension reduction techniques for the free energy landscapes of peptide folding [16]. Although these results show that the intrinsic dimensionality of the dataset may be higher than m = 2, we proceed with our comparative analysis to reductions of two dimensions since this is the dimensionality at which the majority of users are interested in considering the data.…”

Section: Residual Variancementioning

confidence: 99%

Comparative Analysis of Linear and Nonlinear Dimension Reduction Techniques on Mass Cytometry Data

Konstorum

Vidal

Jekel

et al. 2018

Preprint

View full text Add to dashboard Cite

Mass cytometry, also known as CyTOF, is a newly developed technology for quantification and classification of immune cells that can allow for analysis of over three dozen protein markers per cell. The high dimensional data that is generated requires innovative methods for analysis and visualization. We conducted a comparative analysis of four dimension reduction techniques -principal component analysis (PCA), isometric feature mapping (Isomap), t-distributed stochastic neighbor embedding (t-SNE), and Diffusion Maps by implementing them on benchmark mass cytometry data sets. We compare the results of these reductions using computation time, residual variance, a newly developed comparison metric we term neighborhood proportion error (NPE), and two-dimensional visualizations. We find that t-SNE and Diffusion Maps are the two most effective methods for preserving relationships of interest among cells and providing informative visualizations. In low dimensional embeddings, t-SNE exhibits well-defined phenotypic clustering. Additionally, Diffusion Maps can represent cell differentiation pathways with long projections along each diffusion component. We thus recommend a complementary approach using t-SNE and Diffusion Maps in order to extract diverse and informative cell relationship information in a two-dimensional setting from CyTOF data.

show abstract

Evaluation of Dimensionality-Reduction Methods from Peptide Folding–Unfolding Simulations

Cited by 34 publications

References 28 publications

Graph representation of protein free energy landscape

Graph representation of protein free energy landscape

Visual Analytics for Deep Embeddings of Large Scale Molecular Dynamics Simulations

Comparative Analysis of Linear and Nonlinear Dimension Reduction Techniques on Mass Cytometry Data

Contact Info

Product

Resources

About