Health Hack 2015

• Our website

This is our server

ssh $USER@146.118.98.44

Our whiteborad plan

Aims:

• To write file parsers to parse fastqc, vcf and coverage files
• To write parsed output into mongoDB
• To query the database to extract the relevant info
• To visualise relevant information

Instructions for creating and using mongoDB

from pymongo improt MongoClient

connection = MongoClient("mongo://localhost:27017")

db_test = connection.healthhack.test

test_dict = {}

db_test.insert(test_dict)

results = db_test.find()

for result in results:
    print result

library(rmongodb)
query <- mongo.bson.from.list(list('city' = 'COLORADO CITY'))
mongo <- mongo.create()
mongo.get.databases(mongo)

mongo.get.database.collections(mongo, db)
db <-  "healthhack"

res <- mongo.distinct(mongo, "healthhack.test_qc" , "uid")

test2 <- mongo.find.all(mongo,  "healthhack.test_qc", list('uid' = "sdat89"))

pops1 <- mongo.find.all(mongo, "healthhack.test_qc", query = list('uid'  = "sdat89", "quality" = list('$gte' = 32)))

test_qc <- mongo.find.all(mongo, "healthhack.test_qc")

Git how to

1. fork https://github.com/serine/HeHackSheHack this repo to your github
2. git clone your version https://github.com/YOURUserName/HeHackSheHack.git
3. don't delete any files just add your code files to that repo
4. git add your files git commit -m your files
5. git push your files to your own repo
6. send pull request

Hosting files from the server

If you like to host html or any other files put them into ~/www directory and access it as such http://146.118.98.44/home/$USER/yourFile.html

General notes on the projects

• have individual track per patient
• show individual patient, but be able to take a subset of patients, for example ethnicity condition and/or other condition type
• make toggle box to see just a caucasian or asians
• have an average in the y-dimension may also consider average in the x-dimension
• have per gene colour block where you would show all possible variance, summarise all variance in that regions, for example have a heatmap cell, use heatmap colours
• bases shouldn't have colour - general notes
• if a block is more ered than in that gene (block) there are more mutations in that gene/block
• mutations burden - similar just a nubmer of mutation in that regions but also weighted
• base quality per experiment and is the least of our interest right now

Sticky notes notes

Take home message

two type of group

1. pathologiest - don't care just want to see the result
2. whereas developers they are interested what have you learn and technologies what are the pros and cons