Global Illumination for Fun and Profit

2025-05-06 0 0 410.49KB 4 页 10玖币

侵权投诉

ProtoFold Neighborhood Inspector

Nicolas F. Chaves-de-Plaza*, Klaus Hildebrandt* and Anna Vilanova**

*CGV group, TU Delft | **Mathematics and Computer Science, TU Eindhoven

1 INTRODUCTION

Post-translational modifications (PTMs) affecting a protein's

residues (amino acids) can disturb its function, leading to illness.

Whether or not a PTM is pathogenic depends on its type and the

status of neighboring residues. In this paper, we present the

ProtoFold Neighborhood Inspector (PFNI), a visualization system

for analyzing residues neighborhoods. The main contribution is a

visualization idiom, the Residue Constellation (RC), for identifying

and comparing three-dimensional neighborhoods based on per-

residue features and spatial characteristics. The RC leverages two-

dimensional representations of the protein's three-dimensional

structure to overcome problems like occlusion, easing the analysis

of neighborhoods that often have complicated spatial arrangements.

Using the PFNI, we explored proteins’ structural PTM data, which

allowed us to identify patterns in the distribution and quantity of

per-neighborhood PTMs that might be related to their pathogenic

status. In the following, we define the tasks that guided the

development of the PFNI and describe the data sources we derived

and used. Then, we introduce the PFNI and illustrate its usage

through an example of an analysis workflow. We conclude by

reflecting on preliminary findings obtained while using the tool on

the provided data and future directions concerning the development

of the PFNI.

2 DATA

The challenge organizers provided two datasets that describe how

PTMs relate to proteins’ structures1. First, the structures dataset

includes, for each residue (amino acid) of three proteins across two

organisms, the 3D coordinates of its alpha, beta, and carboxyl

carbon atoms and its nitrogen atom. It also includes, for each

residue, an estimate of the prediction confidence (pLLDT) and the

type of structure. Second, the modifications dataset lists the PTMs

that affect the residues in the structures dataset. Each row in the

modifications dataset describes a PTM that affects a protein´s

residue, including the PTM’s name/type, classification, and

pathogenic status. A residue in the structures dataset can be related

to multiple modifications.

Finally, we derived a third dataset, neighbors, from the structures

dataset. The neighbors dataset lists the thirty nearest neighbors of

every residue within every protein structure. We used the Euclidean

distance between the residues’ alpha carbons as the metric to filter

the neighbors.

3 MODELING AND TASKS

We define the general goal of the visualization system as enabling

the analysis of three-dimensional neighborhoods of proteins'

residues. Before proceeding, we formulate this goal explicitly using

the language of graph theory.

We model proteins as graphs where residues are nodes indexed

by the position in the protein's primary structure. To simplify

modeling, we use the position of the residue’s alpha carbon as a

proxy of the residue’s position. In terms of links between nodes,

there are two types: topological links between consecutive residues

and spatial ones between residues that are close in three-

dimensional space. The PFNI focuses on the latter. Finally, nodes

and links have attributes. Examples of the former include the

number and type of PTMs and the pLLDT score. As for the latter,

we consider the Euclidean distance between the nodes that the link

connects.

Considering this definition of the problem and the overall goal,

we now list the tasks that the PFNI should support:

• (T1) Select nodes (residues) for neighborhood analysis

based on node-level and neighborhood-level characteristics.

• (T2) Identify salient patterns in the distribution of a

neighborhood's node and link-level characteristics.

• (T3) Compare neighborhoods to assess their similarity, or

dissimilarity, under a feature of interest.

4 VISUALIZATION SYSTEM: PFNI

Figure 1 presents an overview of the ProtoFold Neighborhood

Inspector (PFNI). We developed the prototype using web

technologies like D3 [1] and Three.js [2]. The following paragraphs

describe the system’s main idiom, the Residue Constellation, and

the supporting Three-Dimensional and Bulk Selection widgets.

4.1 Residue Constellation

4.1.1 Primary Structure Orbits (PSO)

We use a radial layout to organize the protein’s two-dimensional

primary structure. The main goal of the PSO is to let users identify

residues for subsequent neighborhood-level analysis. The PSO

consists of three orbits, each consisting of as many arcs of equal

length as residues in the protein. Figure 2 presents the three orbits

in detail. The inner orbit displays the position of each residue in the

chain using a sequential color scale. In the middle orbit, the colors

of the arcs change depending on the user-selected categorical

variable of interest. Finally, the outer orbit uses the arcs’ thickness

to encode a user-selected numeric variable. In Figure 2, the numeric

variable is the number of modifications each residue has.

Figure 5 shows an additional layer of information that this layer

provides. When the user hovers over a residue, ribbons connecting

the residue to its neighbors appear. Using this transient ribbon view,

the user can get a sense of the residue’s neighborhood before adding

it to the Neighborhoods Force Layout.

We chose a radial layout to depict the protein’s primary structure

for two reasons. First, it allows positioning residues in large chains

side-by-side, which permits using the orbits’ arc’s thickness to

judge relationships between numerical values. Second, because the

center of the PSO remains empty, it allows an efficient overview of

the neighborhoods via the ribbon view.

4.1.2 Neighborhoods Force Layout (NFL)

In the center of the Residue Constellation lies the Neighborhoods

Force Layout (NFL), which provides a 2D representation of the

protein 3D neighborhoods. Explicitly, this view organizes user-

selected residues (primary nodes) and their neighbors (secondary

nodes) using a force layout, similarly to [3]. The NFL depicts

primary and secondary nodes differently to reduce clutter and allow

multiple levels of neighborhood analysis, as Figures 3 and 4

illustrate. On the one hand, the NFL uses the neighborhood

summarization glyph (following subsection) to depict primary

Contact: n.f.chavesdeplaza@tudelft.nl

1 http://biovis.net/2022/biovisChallenges_vis/

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

ProtoFoldNeighborhoodInspectorNicolasF.Chaves-de-Plaza*,KlausHildebrandt*andAnnaVilanova***CGVgroup,TUDelft|**MathematicsandComputerScience,TUEindhoven1INTRODUCTIONPost-translationalmodifications(PTMs)affectingaprotein'sresidues(aminoacids)candisturbitsfunction,leadingtoillness.WhetherornotaPTMispat...

展开>> 收起<<

Global Illumination for Fun and Profit.pdf

共4页,预览1页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

Global Illumination for Fun and Profit

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: