Error: loading failed - Githubissues

Wanted to take a look at MSGFgui but ran into some trouble during install of MSGFplus. See below. I installed the latest Java runtime but get the same error.

install_github('MSGFplus', 'thomasp85') Downloading github repo thomasp85/MSGFplus@master Installing MSGFplus '/Library/Frameworks/R.framework/Resources/bin/R' --vanilla CMD INSTALL \ '/private/var/folders/n4/rlnqcs4101b9m602tytx0kbc0000gn/T/RtmpwIvewh/devtools23e51698cc9/thomasp85-MSGFplus-bb3a656' \ --library='/Users/vnijs/Library/R/3.1/library' --install-tests

installing source package ‘MSGFplus’ ... * R * inst * tests * preparing package for lazy loading \ help ** installing help indices * building package indices * installing vignettes * testing if installed package can be loaded Warning: running command ''java' -version 2>&1' had status 1 Error : .onLoad failed in loadNamespace() for 'MSGFplus', details: call: if (as.numeric(sub(".\"\d.(\d).", "\1", javaVersion)) < error: argument is of length zero Error: loading failed Execution halted ERROR: loading failed

removing ‘/Users/vnijs/Library/R/3.1/library/MSGFplus’ Error: Command failed (1) In addition: Warning message: Username parameter is deprecated. Please use thomasp85/MSGFplus

First of all - can I get you to install through Bioconductor? The stable version is part of that…

Which system are you running on?

What do you get if you type "java -version” in the terminal/cmd?

On 14 Dec 2014, at 16:45, Vincent Nijs notifications@github.com wrote:

Wanted to take a look at MSGFgui but ran into some trouble during install of MSGFplus. See below. I installed the latest Java runtime but get the same error.

install_github('MSGFplus', 'thomasp85') Downloading github repo thomasp85/MSGFplus@master Installing MSGFplus '/Library/Frameworks/R.framework/Resources/bin/R' --vanilla CMD INSTALL \ '/private/var/folders/n4/rlnqcs4101b9m602tytx0kbc0000gn/T/RtmpwIvewh/devtools23e51698cc9/thomasp85-MSGFplus-bb3a656' \ --library='/Users/vnijs/Library/R/3.1/library' --install-tests

installing source package ‘MSGFplus’ ... * R * inst * tests * preparing package for lazy loading \ help ** installing help indices * building package indices * installing vignettes * testing if installed package can be loaded Warning: running command ''java' -version 2>&1' had status 1 Error : .onLoad failed in loadNamespace() for 'MSGFplus', details: call: if (as.numeric(sub(".\"\d.(\d).", "\1", javaVersion)) < error: argument is of length zero Error: loading failed Execution halted ERROR: loading failed removing ‘/Users/vnijs/Library/R/3.1/library/MSGFplus’ Error: Command failed (1) In addition: Warning message: Username parameter is deprecated. Please use thomasp85/MSGFplus — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30.

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out.

I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository.

The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out...

To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : )

You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)...

Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can…

Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology…

best

Thomas

Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Thomas,

The combination of R and javascript is just great. A colleague of mine was a software developer and suggested to use an MVC type approach for the app with HTML templates. I like that idea but since Shiny apps are really one page apps I am not sure if you could replicate the UI dynamics with html easily while the settings from one part of the app are not forgotten when you move to another part. Thoughts?

I tried MSGFgui again today to see if the issue I was having still occurred. Last time I could only navigate one level up or down. For some reason it worked fine today. I assume that means I did something weird last time. If the issue pops up again I will document it and let you know.

I would love to know what resources you found most useful in learning html/css/javascript/d3. At the moment it looks like several of my interests in visualization might be addressed through ggvis and http://www.r-bloggers.com/htmlwidgets-javascript-data-visualization-for-r/. It is mainly creating UI elements/inputs that I would really like to know more about at the moment.

For example, the settings modal in MSGFgui looks really nice. Easily adding inputs like would be great. Shiny has some of that as well but how do you add a button to sort for each new variable? Can a user of MSGFgui save and reuse their inputs from the settings modal?

Another example, would be to create a gui for data-munging with dplyr. Joe Cheng put https://github.com/jcheng5/pipecomposer/blob/master/www/pipecomposer.js on github that is a nice start. But how would you add keyboard shortcut, how would you dynamically add input fields, suggest variables to use, etc., etc.

Another example, would be to create a table view of data with sorting (like Data.Tables) but then add a filter option for each column just above/below that column. E.g., radio buttons for a factor, a slider for dates, etc. Perhaps also adding a text input for a variable drop down where you could typo '> 5000' for the selected variable. Then when the user has indicated how they would like to filter the data, provide an option to save the filter, perhaps as an attribute of the data.frame. That way when the user opens the data.frame again there could be a dropdown with relevant filters.

The same things with plots. dygraphs has some nice options to select data-ranges for time series. NVD3 has nice options where you click an element of the legend to show only part of the data http://shiny.rstudio.com/gallery/nvd3-line-chart-output.html. Etc. etc. How would you add something like that to, e.g., a legend in a ggvis plot? Lots of ideas but not the skills (or time) to implement them :)

I am planning to reorganize the code for my app so that some of the functions to summarize and plot data are exported and could be called from, for example, an R-markdown file to reproduce the research results. Wouldn't it be great if you could access the R process that shiny is running, e.g., to call the data in memory? I noticed you have a function to allow data access. I am going to try something like https://gist.github.com/ptoche/8405209 to 'pause' the shiny app and query (and perhaps change) elements of a reactiveValues object that would then be used by the Shiny app after you un-pause the app. We'll see if that works. Anyway, I will use the structure of PanVizGenerator as a template. Thank you.

I asked a silly question about shinyFiles and then stumbled onto your other repo's. I am broadly interested in data-science for business but various fields, such as computational biology, use some of the same tools (e.g., MDS, PCA, clustering).

Vincent

— Sent from Mailbox

On Wed, Dec 17, 2014 at 12:11 AM, Thomas Lin Pedersen notifications@github.com wrote:

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out. I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository. The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out... To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : ) You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)... Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can… Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology… best Thomas Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67289423

Hi,

I got the same error on my linux machine. My java version is:

    java -version
    openjdk version "1.8.0_25"
    OpenJDK Runtime Environment (build 1.8.0_25-b18)
    OpenJDK 64-Bit Server VM (build 25.25-b02, mixed mode)

The error is:

* installing *source* package ‘MSGFplus’ ...
** R
** inst
** preparing package for lazy loading
** help
*** installing help indices
** building package indices
** installing vignettes
** testing if installed package can be loaded
Error : .onLoad failed in loadNamespace() for 'MSGFplus', details:
  call: if (as.numeric(sub(".*\"\\d\\.(\\d).*", "\\1", javaVersion)) < 
  error: argument is of length zero
Error: loading failed
Execution halted
    ERROR: loading failed

Greetz

Hi Vincent

I just had time to look into your Radiant app. It is really impressive - especially the state saving and markdown support. I’ll probably steal those implementations for some of my future projects :-)

good job!

Thomas

On 21 Dec 2014, at 05:05, Vincent Nijs notifications@github.com wrote:

Thomas,

The combination of R and javascript is just great. A colleague of mine was a software developer and suggested to use an MVC type approach for the app with HTML templates. I like that idea but since Shiny apps are really one page apps I am not sure if you could replicate the UI dynamics with html easily while the settings from one part of the app are not forgotten when you move to another part. Thoughts?

I tried MSGFgui again today to see if the issue I was having still occurred. Last time I could only navigate one level up or down. For some reason it worked fine today. I assume that means I did something weird last time. If the issue pops up again I will document it and let you know.

I would love to know what resources you found most useful in learning html/css/javascript/d3. At the moment it looks like several of my interests in visualization might be addressed through ggvis and http://www.r-bloggers.com/htmlwidgets-javascript-data-visualization-for-r/. It is mainly creating UI elements/inputs that I would really like to know more about at the moment.

For example, the settings modal in MSGFgui looks really nice. Easily adding inputs like would be great. Shiny has some of that as well but how do you add a button to sort for each new variable? Can a user of MSGFgui save and reuse their inputs from the settings modal?

Another example, would be to create a gui for data-munging with dplyr. Joe Cheng put https://github.com/jcheng5/pipecomposer/blob/master/www/pipecomposer.js on github that is a nice start. But how would you add keyboard shortcut, how would you dynamically add input fields, suggest variables to use, etc., etc.

Another example, would be to create a table view of data with sorting (like Data.Tables) but then add a filter option for each column just above/below that column. E.g., radio buttons for a factor, a slider for dates, etc. Perhaps also adding a text input for a variable drop down where you could typo '> 5000' for the selected variable. Then when the user has indicated how they would like to filter the data, provide an option to save the filter, perhaps as an attribute of the data.frame. That way when the user opens the data.frame again there could be a dropdown with relevant filters.

The same things with plots. dygraphs has some nice options to select data-ranges for time series. NVD3 has nice options where you click an element of the legend to show only part of the data http://shiny.rstudio.com/gallery/nvd3-line-chart-output.html. Etc. etc. How would you add something like that to, e.g., a legend in a ggvis plot? Lots of ideas but not the skills (or time) to implement them :)

I am planning to reorganize the code for my app so that some of the functions to summarize and plot data are exported and could be called from, for example, an R-markdown file to reproduce the research results. Wouldn't it be great if you could access the R process that shiny is running, e.g., to call the data in memory? I noticed you have a function to allow data access. I am going to try something like https://gist.github.com/ptoche/8405209 to 'pause' the shiny app and query (and perhaps change) elements of a reactiveValues object that would then be used by the Shiny app after you un-pause the app. We'll see if that works. Anyway, I will use the structure of PanVizGenerator as a template. Thank you.

I asked a silly question about shinyFiles and then stumbled onto your other repo's. I am broadly interested in data-science for business but various fields, such as computational biology, use some of the same tools (e.g., MDS, PCA, clustering).

Vincent

— Sent from Mailbox

On Wed, Dec 17, 2014 at 12:11 AM, Thomas Lin Pedersen notifications@github.com wrote:

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out. I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository. The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out... To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : ) You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)... Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can… Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology… best Thomas Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67289423 — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67759428.

Thanks Thomas. The calling the report/markdown-creation functions is a bit messy. Hope to clean that up a bit. Suggestions welcome.

For saving state also take a look at https://github.com/jcheng5/shiny-resume/pull/1 This makes saving state easier when running on a server with multiple users. If you have any ideas on how to make this work well with and without renderUI let us know. It doesn't affect me all that much because I only have renderUI in radiant.

On a somewhat related note ... do you have any comments on this question I posted a little while back? Hadley's comment is, I assume, a lot closer to your research area than it is to mine.

Vincent

Sorry for not getting back to you on your SO question - I’ve read through it and I’m afraid I’m not quite sure what the actual question is… How to best handle huge objects in R, or how to best use pass-by-reference semantics in R? If you clarify I’ll be happy to weigh in…

On a completely different note (and related to one of your comments in this thread) I’m working on a more general server-side framework for the occasions where Shiny’s model of coupling server and UI just feels too constrained. The package (Fiery) is only concerned with server-side logic, making it possible to do whatever on the front-end you may wish (including using a variety of front-end MVC frameworks such as Angular and Ember). It is in the very early stage at the moment, and is a pass time project, but I thought you may be interested…

/Thomas

On 21 Dec 2014, at 05:05, Vincent Nijs notifications@github.com wrote:

Thomas,

The combination of R and javascript is just great. A colleague of mine was a software developer and suggested to use an MVC type approach for the app with HTML templates. I like that idea but since Shiny apps are really one page apps I am not sure if you could replicate the UI dynamics with html easily while the settings from one part of the app are not forgotten when you move to another part. Thoughts?

I tried MSGFgui again today to see if the issue I was having still occurred. Last time I could only navigate one level up or down. For some reason it worked fine today. I assume that means I did something weird last time. If the issue pops up again I will document it and let you know.

I would love to know what resources you found most useful in learning html/css/javascript/d3. At the moment it looks like several of my interests in visualization might be addressed through ggvis and http://www.r-bloggers.com/htmlwidgets-javascript-data-visualization-for-r/. It is mainly creating UI elements/inputs that I would really like to know more about at the moment.

For example, the settings modal in MSGFgui looks really nice. Easily adding inputs like would be great. Shiny has some of that as well but how do you add a button to sort for each new variable? Can a user of MSGFgui save and reuse their inputs from the settings modal?

Another example, would be to create a gui for data-munging with dplyr. Joe Cheng put https://github.com/jcheng5/pipecomposer/blob/master/www/pipecomposer.js on github that is a nice start. But how would you add keyboard shortcut, how would you dynamically add input fields, suggest variables to use, etc., etc.

Another example, would be to create a table view of data with sorting (like Data.Tables) but then add a filter option for each column just above/below that column. E.g., radio buttons for a factor, a slider for dates, etc. Perhaps also adding a text input for a variable drop down where you could typo '> 5000' for the selected variable. Then when the user has indicated how they would like to filter the data, provide an option to save the filter, perhaps as an attribute of the data.frame. That way when the user opens the data.frame again there could be a dropdown with relevant filters.

The same things with plots. dygraphs has some nice options to select data-ranges for time series. NVD3 has nice options where you click an element of the legend to show only part of the data http://shiny.rstudio.com/gallery/nvd3-line-chart-output.html. Etc. etc. How would you add something like that to, e.g., a legend in a ggvis plot? Lots of ideas but not the skills (or time) to implement them :)

I am planning to reorganize the code for my app so that some of the functions to summarize and plot data are exported and could be called from, for example, an R-markdown file to reproduce the research results. Wouldn't it be great if you could access the R process that shiny is running, e.g., to call the data in memory? I noticed you have a function to allow data access. I am going to try something like https://gist.github.com/ptoche/8405209 to 'pause' the shiny app and query (and perhaps change) elements of a reactiveValues object that would then be used by the Shiny app after you un-pause the app. We'll see if that works. Anyway, I will use the structure of PanVizGenerator as a template. Thank you.

I asked a silly question about shinyFiles and then stumbled onto your other repo's. I am broadly interested in data-science for business but various fields, such as computational biology, use some of the same tools (e.g., MDS, PCA, clustering).

Vincent

— Sent from Mailbox

On Wed, Dec 17, 2014 at 12:11 AM, Thomas Lin Pedersen notifications@github.com wrote:

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out. I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository. The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out... To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : ) You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)... Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can… Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology… best Thomas Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67289423 — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67759428.

Thomas,

The question refers to a statement Hadley made in his book about how in your area packages apparantly use environments to pass around data. I assume that means that rather than pass an actual data.frame to functions a reference to that data is provided. I was just wondering if you know any examples that demonstrate how to set that up and, perhaps more importantly, how to check that you are doing it correctly. I currently put data in my shiny app in a reactive value which is a special kind of list. I don't know if there is something like a reactive environment in Shiny. There might be. Rather than pass around data I pass a string with the dataset name. The data is then accessed by looking up the appropriate entry in the reactiveValues object (r_value in my app). Comments/suggestions welcome.

Fiery sounds interesting! Might be a challenge to 'compete' with the shiny framework though. There is also opencpu which does, I think, something similar that what you are thinking about. It does allow you to use Angular and others as a front end. Jeroen Ooms has a bunch of JavaScript stuff that may be of some relevance for you (e.g., js and V8). The htmlwidgets package also allows communication between R and js-libraries.

You may also be interested in Spyre. There are a few projects on github with that name so you would have to check which is the R based one. I seem to recall there was also a project that does something similar to Shiny but with a different type of reactive model. Can't remember the name but I know it is on github.

I am not an expert in server-side tech but I am not sure using R as the server is likely to produce a high performance system. What about a full-stack Java-script system (e.g., meteor) that calls R only for analysis as needed?

I actually really like Shiny and wouldn't have been able to create radiant without it. My current knowledge of JavaScript is limited so easy access for a regular R user would be important. That said, my main pain-points in Shiny are:

Difficult to reuse code because inputs cannot be reused. This seems to be a javascript issue though.
Testing reactive functions and determining where an error occurs. Finding bugs in Python seemed much easier than in R in general however.
Shiny's environment is not easily accessible to package function in the R/ directory
Not designed to support package creation. This doesn't help with automated testing either since tools are setup to test code in R (mainly).

Vincent

On Sun, Apr 19, 2015 at 11:58 PM, Thomas Lin Pedersen notifications@github.com wrote:

Vincent Sorry for not getting back to you on your SO question - I’ve read through it and I’m afraid I’m not quite sure what the actual question is… How to best handle huge objects in R, or how to best use pass-by-reference semantics in R? If you clarify I’ll be happy to weigh in… On a completely different note (and related to one of your comments in this thread) I’m working on a more general server-side framework for the occasions where Shiny’s model of coupling server and UI just feels too constrained. The package (Fiery) is only concerned with server-side logic, making it possible to do whatever on the front-end you may wish (including using a variety of front-end MVC frameworks such as Angular and Ember). It is in the very early stage at the moment, and is a pass time project, but I thought you may be interested… /Thomas

On 21 Dec 2014, at 05:05, Vincent Nijs notifications@github.com wrote:

Thomas,

The combination of R and javascript is just great. A colleague of mine was a software developer and suggested to use an MVC type approach for the app with HTML templates. I like that idea but since Shiny apps are really one page apps I am not sure if you could replicate the UI dynamics with html easily while the settings from one part of the app are not forgotten when you move to another part. Thoughts?

I tried MSGFgui again today to see if the issue I was having still occurred. Last time I could only navigate one level up or down. For some reason it worked fine today. I assume that means I did something weird last time. If the issue pops up again I will document it and let you know.

I would love to know what resources you found most useful in learning html/css/javascript/d3. At the moment it looks like several of my interests in visualization might be addressed through ggvis and http://www.r-bloggers.com/htmlwidgets-javascript-data-visualization-for-r/. It is mainly creating UI elements/inputs that I would really like to know more about at the moment.

For example, the settings modal in MSGFgui looks really nice. Easily adding inputs like would be great. Shiny has some of that as well but how do you add a button to sort for each new variable? Can a user of MSGFgui save and reuse their inputs from the settings modal?

Another example, would be to create a gui for data-munging with dplyr. Joe Cheng put https://github.com/jcheng5/pipecomposer/blob/master/www/pipecomposer.js on github that is a nice start. But how would you add keyboard shortcut, how would you dynamically add input fields, suggest variables to use, etc., etc.

Another example, would be to create a table view of data with sorting (like Data.Tables) but then add a filter option for each column just above/below that column. E.g., radio buttons for a factor, a slider for dates, etc. Perhaps also adding a text input for a variable drop down where you could typo '> 5000' for the selected variable. Then when the user has indicated how they would like to filter the data, provide an option to save the filter, perhaps as an attribute of the data.frame. That way when the user opens the data.frame again there could be a dropdown with relevant filters.

The same things with plots. dygraphs has some nice options to select data-ranges for time series. NVD3 has nice options where you click an element of the legend to show only part of the data http://shiny.rstudio.com/gallery/nvd3-line-chart-output.html. Etc. etc. How would you add something like that to, e.g., a legend in a ggvis plot? Lots of ideas but not the skills (or time) to implement them :)

I am planning to reorganize the code for my app so that some of the functions to summarize and plot data are exported and could be called from, for example, an R-markdown file to reproduce the research results. Wouldn't it be great if you could access the R process that shiny is running, e.g., to call the data in memory? I noticed you have a function to allow data access. I am going to try something like https://gist.github.com/ptoche/8405209 to 'pause' the shiny app and query (and perhaps change) elements of a reactiveValues object that would then be used by the Shiny app after you un-pause the app. We'll see if that works. Anyway, I will use the structure of PanVizGenerator as a template. Thank you.

I asked a silly question about shinyFiles and then stumbled onto your other repo's. I am broadly interested in data-science for business but various fields, such as computational biology, use some of the same tools (e.g., MDS, PCA, clustering).

Vincent

— Sent from Mailbox

On Wed, Dec 17, 2014 at 12:11 AM, Thomas Lin Pedersen notifications@github.com wrote:

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out. I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository. The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out... To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : ) You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)... Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can… Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology… best Thomas Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67289423 — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67759428.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-94374511

Ok. I think I get the question now. Of course I can’t really speak for what Hadley were referring to but once upon a time this was more relevant than it is now. For an example you could have a look at xcms and the xcmsRaw class therein (it’s a bioconductor package). With that being set I don’t really see much point in replicating that behaviour anymore - while R uses pass-by-copy, it uses it in a smart way so that elements are only copied once they are modified. Passing around a large object in R is thus not a problem as long as you only read from it. The main selling point of environments instead of lists for huge datasets was because lists would make a complete deep copy every time an element was changed - quite a costly operation! Environments didn’t have this problem so they were more efficient for storing data that continuously needed to be modified. This all changed in R 3.1 (or somewhere around that version) where lists were changed so they avoided the deep copy of the whole list on element mutations…

Bottom line is that environments now are mostly useful if you want to program in a more C++/javascript/python kind of object oriented way, with reference passing and methods linked to objects instead of generics - this behaviour is implemented in refClasses and the R6 package for instance.

To sum up, I don’t think you need to worry so much about how you store your data within Shiny. Of course if you have really huge data, and want persistence then it should be put in a database, but otherwise just use lists or reactiveValues…

As for Fiery, I have no intentions of competing with Shiny - I even state it clearly that I believe Shiny would be the right choice in 99% of all use cases - I’ll look in to some of the packages you mentioned (those I don’t know already). Your concern for the suitability of R as a backend engine is valid, but in my case the infrastructure offered by Bioconductor is unparalleled within bioinformatics. For these types of applications (such as MSGFgui), where R is mostly used as a data store/analysis engine it makes good sense. But you’re right that whatever Fiery ends up being, it will probably never compete with Django or Ruby on Rails… (nor should it)

best

Thomas

On 20 Apr 2015, at 19:45, Vincent Nijs notifications@github.com wrote:

Thomas,

The question refers to a statement Hadley made in his book about how in your area packages apparantly use environments to pass around data. I assume that means that rather than pass an actual data.frame to functions a reference to that data is provided. I was just wondering if you know any examples that demonstrate how to set that up and, perhaps more importantly, how to check that you are doing it correctly. I currently put data in my shiny app in a reactive value which is a special kind of list. I don't know if there is something like a reactive environment in Shiny. There might be. Rather than pass around data I pass a string with the dataset name. The data is then accessed by looking up the appropriate entry in the reactiveValues object (r_value in my app). Comments/suggestions welcome.

Fiery sounds interesting! Might be a challenge to 'compete' with the shiny framework though. There is also opencpu which does, I think, something similar that what you are thinking about. It does allow you to use Angular and others as a front end. Jeroen Ooms has a bunch of JavaScript stuff that may be of some relevance for you (e.g., js and V8). The htmlwidgets package also allows communication between R and js-libraries.

You may also be interested in Spyre. There are a few projects on github with that name so you would have to check which is the R based one. I seem to recall there was also a project that does something similar to Shiny but with a different type of reactive model. Can't remember the name but I know it is on github.

I am not an expert in server-side tech but I am not sure using R as the server is likely to produce a high performance system. What about a full-stack Java-script system (e.g., meteor) that calls R only for analysis as needed?

I actually really like Shiny and wouldn't have been able to create radiant without it. My current knowledge of JavaScript is limited so easy access for a regular R user would be important. That said, my main pain-points in Shiny are:

Difficult to reuse code because inputs cannot be reused. This seems to be a javascript issue though.

Testing reactive functions and determining where an error occurs. Finding bugs in Python seemed much easier than in R in general however.

Shiny's environment is not easily accessible to package function in the R/ directory

Not designed to support package creation. This doesn't help with automated testing either since tools are setup to test code in R (mainly).

Vincent

On Sun, Apr 19, 2015 at 11:58 PM, Thomas Lin Pedersen notifications@github.com wrote:

Vincent Sorry for not getting back to you on your SO question - I’ve read through it and I’m afraid I’m not quite sure what the actual question is… How to best handle huge objects in R, or how to best use pass-by-reference semantics in R? If you clarify I’ll be happy to weigh in… On a completely different note (and related to one of your comments in this thread) I’m working on a more general server-side framework for the occasions where Shiny’s model of coupling server and UI just feels too constrained. The package (Fiery) is only concerned with server-side logic, making it possible to do whatever on the front-end you may wish (including using a variety of front-end MVC frameworks such as Angular and Ember). It is in the very early stage at the moment, and is a pass time project, but I thought you may be interested… /Thomas

On 21 Dec 2014, at 05:05, Vincent Nijs notifications@github.com wrote:

Thomas,

The combination of R and javascript is just great. A colleague of mine was a software developer and suggested to use an MVC type approach for the app with HTML templates. I like that idea but since Shiny apps are really one page apps I am not sure if you could replicate the UI dynamics with html easily while the settings from one part of the app are not forgotten when you move to another part. Thoughts?

I tried MSGFgui again today to see if the issue I was having still occurred. Last time I could only navigate one level up or down. For some reason it worked fine today. I assume that means I did something weird last time. If the issue pops up again I will document it and let you know.

I would love to know what resources you found most useful in learning html/css/javascript/d3. At the moment it looks like several of my interests in visualization might be addressed through ggvis and http://www.r-bloggers.com/htmlwidgets-javascript-data-visualization-for-r/. It is mainly creating UI elements/inputs that I would really like to know more about at the moment.

For example, the settings modal in MSGFgui looks really nice. Easily adding inputs like would be great. Shiny has some of that as well but how do you add a button to sort for each new variable? Can a user of MSGFgui save and reuse their inputs from the settings modal?

Another example, would be to create a gui for data-munging with dplyr. Joe Cheng put https://github.com/jcheng5/pipecomposer/blob/master/www/pipecomposer.js on github that is a nice start. But how would you add keyboard shortcut, how would you dynamically add input fields, suggest variables to use, etc., etc.

Another example, would be to create a table view of data with sorting (like Data.Tables) but then add a filter option for each column just above/below that column. E.g., radio buttons for a factor, a slider for dates, etc. Perhaps also adding a text input for a variable drop down where you could typo '> 5000' for the selected variable. Then when the user has indicated how they would like to filter the data, provide an option to save the filter, perhaps as an attribute of the data.frame. That way when the user opens the data.frame again there could be a dropdown with relevant filters.

The same things with plots. dygraphs has some nice options to select data-ranges for time series. NVD3 has nice options where you click an element of the legend to show only part of the data http://shiny.rstudio.com/gallery/nvd3-line-chart-output.html. Etc. etc. How would you add something like that to, e.g., a legend in a ggvis plot? Lots of ideas but not the skills (or time) to implement them :)

I am planning to reorganize the code for my app so that some of the functions to summarize and plot data are exported and could be called from, for example, an R-markdown file to reproduce the research results. Wouldn't it be great if you could access the R process that shiny is running, e.g., to call the data in memory? I noticed you have a function to allow data access. I am going to try something like https://gist.github.com/ptoche/8405209 to 'pause' the shiny app and query (and perhaps change) elements of a reactiveValues object that would then be used by the Shiny app after you un-pause the app. We'll see if that works. Anyway, I will use the structure of PanVizGenerator as a template. Thank you.

I asked a silly question about shinyFiles and then stumbled onto your other repo's. I am broadly interested in data-science for business but various fields, such as computational biology, use some of the same tools (e.g., MDS, PCA, clustering).

Vincent

— Sent from Mailbox

On Wed, Dec 17, 2014 at 12:11 AM, Thomas Lin Pedersen notifications@github.com wrote:

On 17 Dec 2014, at 01:12, Vincent Nijs notifications@github.com wrote:

Issue resolved (see below). Is MSGFgui all in Shiny? It looks amazing!!! Is java just for the package that reads the xls files? I got the milk-protien.fasta file from your github site to explore the interface a bit. For some reason I had to put the file in the root directory because the file browser wouldn't let me navigate to my Desktop (Mac OSX Mavericks). Haven't been able to get the app to do any analysis yet but I am not in a bio related field so that is probably my fault because I don't know any of the terminology. Could you point me to an example data file that should work with MSGFgui?

Glad you got it to work and happy that you find it amazing : ). It is a very special blend of custom javascript, css, javascript (with D3) and shiny for communicating back to R. What I mean with this is that I mostly use Shiny for the client-server communication, and rely on standard web technologies for most of what is happening at the UI level. Being mainly interested in data visualisation and data interfaces I find the standard Shiny look too limited and generic (it’s fine for a GUI overlay for simple functionality though). If you want to discuss this approach in more detail then feel free to reach out. I’m baffled by the fact that you couldn’t get to the desktop - could you navigate anywhere else? Otherwise please post an issue on my shinyFiles repository. The raw data files used for the MSGFgui analysis are rather big which is the reason why it isn’t distributed as an example. The results are also fairly difficult to understand without any biological insight. If you are very interested I’ll set up a link where you can download the relevant files to try it out... To be honest your PanViz, from the video you put up, looks even more amazing. Very curious to see that in action. PanViz is all javascript/D3 I assume?

Again - happy you like it. Yes PanViz itself is all JavaScript/D3, while PanVizGenerator is a standard shiny app. It’s supposed to be more of a wow-experience than MSGFgui as it is an exercise in pure interactive data visualisation, while MSGFgui is more concerned with blending standard GUI element with data visualisation based navigation. Furthermore it was what I used as a project to learn html/css/javascript/d3 so I did go overboard some places just for the sake of pushing myself : ) You can try it out by downloading the example data and using that for input. Unpack the zip file you download and open the index.html file in your favourite browser (Chrome and Safari are the most performant)... Your shiny apps are some of the few that I have seen that have an R-package structure so I plan to use PanVizGenerate as a template to structure my own app if you don't mind.

I have struggled with this quite a bit as Shiny doesn’t seem to be too concerned with this aspect. What I have aimed for is to have all R code in the R folder instead of residing in server.R and ui.R (I really don’t use ui.R as I usually work with custom html), while retaining the possibility of uploading the resulting app to shinyapps.io. PanVizGenerator is my best stab at it, but approaches might change as the Shiny devs begin to focus on this aspect (if they ever do that). Feel free to use anything you can… Can I ask, out of curiosity, how you stumbled upon my work? You seem to be mostly into marketing, not computational biology… best Thomas Keep up the amazing work!

The fix you proposed ... I installed:

java version "1.8.0_25" Java(TM) SE Runtime Environment (build 1.8.0_25-b17) Java HotSpot(TM) 64-Bit Server VM (build 25.25-b02, mixed mode)

Then when I started MSFgui after installing from biocLite it asked to install Java runtime 6 :)

So replacing: install_github('shinyFiles', 'thomasp85') install_github('MSGFplus', 'thomasp85') install_github('MSGFgui', 'thomasp85')

With: biocLite('shinyFiles') biocLite('MSGFplus') biocLite('MSGFgui')

Then continue as specified with the below and it starts: library(MSGFgui) MSGFgui()

— Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67257005.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67289423 — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-67759428.

Reply to this email directly or view it on GitHub: https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-94374511 — Reply to this email directly or view it on GitHub https://github.com/thomasp85/MSGFgui/issues/30#issuecomment-94521843.

Thanks for the reference to xcms although it sounds like using environments may not be as influential as I thought. I'll focus on adding interaction with databases first.

thomasp85 / MSGFgui

Error: loading failed #30