23 June 2017 Meeting Minutes
A teleconference/webex was held on 23 June 2017. P Drew gave an update on his attendance at the European PhUSE CSS conference in London during the past week. A nonclinical topics group met at the conference and had a specific session on data visualization. Phil presented our working group's poster from the March 2017 PhUSE CSS with positive feedback. Demos of software from Instem (SEND Explorer), PointCross and Xybion were provided by vendors. Crystal Allard from US FDA discussed use of visualization tools at FDA. FDA using a template approach to simplify software applications. Given that the various visualization tools have many degrees of freedom, group should consider whether guidance on structure/templates for tools should be provided. By limiting the variability of user options, complexity can be decreased. Can use-case examples be developed to provide guidance for tools ? Organize the functionality of the software tools. Example could be predefined parameter settings for displaying/analyzing for liver toxicity. Review of Histographic tool was conducted. A Brown inquired and suggested a proposed hierarchy of data to be presented from inner to outer rings: tissue, finding, dose, period, sex, individual animal. K Snyder confirmed this was indeed the current setting in software. Color range for clinical chemistry values (ALKP) as example based on magnitude of value or fold-change to control: highest is red, lowest blue, intermediate is green. Group discussed identifying a liver panel or a kidney panel for linking clinical pathology parameters to histopathology data and displaying via Histographic tool. Clinical pathology data can be selected and displayed simultaneous with histopathology data, with color codes based on magnitude of change. PointCross will share open source tool for incorporating SEND data for visual display purposes. This tool is in development and group can explore its use. P Marc will distribute another novel visualization tool for the group to evaluate (Circos).
Post meeting note sent by P. Marc
I just wanted to highlight the Circos library used for genomics analysis. The density of information you can obtain is incredible. We could easily pack data from a full study, including histopath and clinpath in a single graph I think.
A good example here: a very large image + an automated “mouse-over” zoom, + links on each pixel to drill down @ https://research.nhgri.nih.gov/bic/circos/BIC_Circos_BRCA1no11-zoomer.shtml
Circos is available for free here: http://circos.ca
“old fashion” and less interactive then Krona, but extremely powerful to pack data in a maximum density!