Object2vr tutorial update#

Some of those edges may be “REF” edges, i.e. connecting to that node’s reference-adjacent neighbor. Using this system of orientation, there is an intrinsic direction associated with every node, i.e. there are edges associated with its “left” and “right” side. In the human genome, the 1q telomere is thus on the “right” side of chromosome 1, and the 1p telomere is on its “left”. As a convention we refer to the direction of increasing coordinates on the reference as “right” and the direction of decreasing coordinates as “left”. Conversely, the “-” strand of the genome goes from 5’ to 3’ in the direction of decreasing genomic coordinates. Just to review, strands of any (double-stranded) reference genome are arbitrarily labeled “+” and “-”, where the “+” strand goes from 5’ to 3’ in the direction of increasing genomic coordinates. We contrast this skew-symmetric structure to the traditional “bidirected graph” formulation of DNA assemblies, which uses nodes to represent double-stranded sequence “ends” and two flavors of (intrasegment, intersegment) edges to represent sequences and their adjacencies, respectively. Though we retain the underlying structure of a directed graph (convenient for applying standard graph algorithms) we are able to model the “sidedness” of DNA sequence. Adjacencies associated with the left side of one node can connect to the left or right side of any other node. However, since gGraph nodes represent interval / anti-interval pairs, we can talk about their “right-side” (i.e. the side with adjacencies leaving the “+” interval and entering the “-” interval) and their “left side” (i.e. the side with adjacencies leaving the “-” interval and entering the “+” interval). Though, in a directed graph, we may arbitrarily label nodes that send edges to our vertex as being to its “left”, and edges that receive edges from our vertex to its “right”, it is impossible to then connect two vertices in our directed graph on their mutual left. This limits the applicability of standard directed graphs to the modeling of DNA, since double-stranded DNA segments can be rearranged with each other in any of four (left/left, right/left, right/right, left/left) possible orientations. The vertices in a standard directed graph do not formally have left and right “sides” that edges can both enter and exit. In this directed graph, every interval has an “anti-interval”, every adjacency has an “anti-adjacency”, and every path (i.e. sequence of intervals or adjacencies) has an “anti-path”. This makes the gGraph a skew-symmetric directed graph. Similarly, every gEdge is a pair of directed adjacencies. Since DNA is double stranded, every gNode in the gGraph is actually a pair of stranded GRanges, one on the “+” and the other on the “-” reference strand.

Object2vr tutorial update#

If you use gGnome in your work, please contact us so that we can update you when the citation becomes available. GGnome is currently in an alpha release and a manuscript in preparation. For background, it may help to have some familiarity with data.table, GenomicRanges, gUtils, igraph packages. However, we believe that the package can be useful in any context where graphs and paths of reference genomic intervals are generated, such as through the analysis of complex germline haplotypes, population genetics of structural variation, partially assembled genome drafts, comparative genomics, and transcriptome assembly.įor installation instructions, please visit the gGnome github page.

Our key interest in developing gGnome is to develop a framework to analyze the sorts of graphs that arise in the study of whole genome cancer structural variation. Please see here for more details about the R6 standard. This includes methods that (similar to other object oriented languages) enable the object to be modified “in place”. Because gGnome classes are written in R6, their methods, variables, and “active bindings” are referenced using the $ symbol.

gGnome is written in the R6 object oriented standard and built around a powerful GenomicRanges, data.table, and igraph backend, and thus supports agile interaction with graphs consisting of hundreds of thousands of nodes and edges. The gGnome package provides a flexible, queriable R interface to graphs and walks of genomic intervals.