Kaitlyn’s notebook: Cheese grater and Java (for a new package BACA)

I setup cheese grater in 213. It is connected to the same mouse/keyboard and monitor switch as Emu, Roadrunner and Ostrich. Cheese grater is #1. I booted it up twice and it worked on the second try. Sam created a srlab account for me to evaluate it a little bit. Cheese grater is running OsX Lion so although it can run code well, it can’t neccessarily use Chrome or Firefox well ebcause support was dropped from it in 2014. For example, Github doesn’t work at all. Therefore, we decided to install Ubuntu 16.04.4 Xenial/Xerus on it. I tried installing it from the USB Sam provided me, however upon rebooting, only the hard drives were avaliable to select rather than giving the USB as an option. Cheese grater wasn’t reading the USB drive at any point in the process even after it was reformatted. The next option was to use a disk to install Ubuntu. Sam entered the same code as before and it was taking significantly longer to run, and hadn’t finished before I left. I will continue working with it tomorrow.

I removed all unexpressed proteins and ran Kmeans clustering on R again for Silo 3. I want to try to ran a for loop in the code however it will also require some if then commands because of teh multiple tables generated. For now, I am focusing on Silo 3.

My ultimate goal is to compare the gene lists from each cluster using DAVID. However, this would leave me with either 25 tables or plots to compare them against. Therefore, I searched for another tool to easily compare gene lists. I found a program called BACA. This looks exactly what I need! It can even generate heatmap plots of multiple gene lists showing similarity, for example (not my data):


BACA requires two packages: ggplot2 and RDAVIDWebService. RDAVIVDWebService would also be a good package to have because I could run gene lists through DAVID with R. However, RDAVIDWebService requires rJava, which is the complicated part of getting these packages to function. As I ahve learned, Java is a nightmare program.

Bioconductor software contains RDAVIDWebService (and another great program, GOsummaries):

## try http:// if https:// URLs are not supported

install.packages("RDAVIDWebService"); library(RDAVIDWebService); library(ggplot2); install.packages("BACA"); library(BACA)

Register with DAVID and begin the tutorial:

result.kegg <- DAVIDsearch(gene.lists.ex, david.user = "###########", 
           idType="ENTREZ_GENE_ID", annotation="KEGG_PATHWAY")

The last line will most likely give you an error:

Error in .jcall(stub, "S", "authenticate", email) : org.apache.axis2.AxisFault: Transport error: 301 Error: Moved Permanently

This is a problem with the certification between Java and DAVID. Therefore, we need to add certification into the keystore and reconfigure Java. Follow these instructions. Unfortunately, these instructions were quite tricky for me. I tried multiple times on Emu with the Ubuntu instructions, but even when the certificate was successfully added to the keystore, Java didn’t reconfigure successfully despite lots of effort…

This problem was really nagging me because I really felt like this plot would be very beneficial to my data. So I grabbed a Windows 10 computer and starting reworking through the new instructions. I had to do several workarounds because of the paths. Cygwin is a bash shell but it still lacks a lot of commands (sudo for example) and paths are different. If you begin following these instructions, be careful to keep track of what you are moving and where things are located.

Ultimately, to get the certificate added to the keystore, I had to move the location of the keytool since it wouldn’t execute outside of the folder is was located in. Therefore my keytool wa located in ../lib/security. Then I had to copy the certificate back to the bin (the original location). I never successfully ran R CMD javarecof (no programs registered the run command). So I moved on to the next code just to see if the new keystore certification was enough. I re ran the DAVID command in R  and received a new error!

Warning message:
In connect() :
Failed. For user registration, go to http://david.abcc.ncifcrf.gov/webservice/register.

I know I registered (both of) my email(s) to DAVID. Honestly, it would be odd if it updated immediately. There is not any mention of this error. Cristol (the package creator) mentions to wait 24-48 hours for registration briefly on a forum with the Java instructions. I’m hoping that will give me access because this was a lot of work, but I think the plots will be worth it! I also want to use GOsummaries which is with the bioconductor software, but I am getting a permission error when installing it. From a few quick google searches it should be a simple fix, but I was more focused on java this time.

![pic](https://d.pr/i/I7TxPE+) leave out TGACCA ACAGTG


leave out TGACCA