CnrLwlss

What makes a great scientific computing language?

• Expressivity: should be possible to express mathematical & programming concepts concisely & elegantly. This is important to reduce the mental overhead required for writing down and reading models and algorithms in the form of code.
• Speed: should be possible to convert code into software/program that gets the most out of modern computer hardware
  • Functional Programming
  • Parallel computing
• Extensibility: should have a powerful, unambiguous, easy to use package management system built right into the core of the language
• Visualisation: a large and very important part of scientific programming is generating side-effects: writing reports, generating plots, creating interactive documents

CnrLwlss

Why Julia?

• Scientific programming language
• Open source
• Package management
• Supports functional programming
• Designed to take advantage of computers with multiple CPUs

Why R?

CnrLwlss

#https://github.com/hrbrmstr/speedtest
library(speedtest)
fname = "speedtest_results2.txt"
makeplots = FALSE

if(!makeplots){
 config = spd_config()
 servers = spd_servers(config = config)
 servers = spd_closest_servers(servers, config)
 best = spd_best_servers(servers, config, max = 3)

CnrLwlss

# Info about boxplot notches https://sites.google.com/site/davidsstatistics/home/notched-box-plots
# Article about why not to use barplots https://doi.org/10.1371/journal.pbio.1002128
# Article about being wary of summary statistics and why raw data plots are better than boxplots https://www.autodeskresearch.com/publications/samestats

# Generate some fake data
concs = seq(0,10,1)
concobs = rep(concs,each=500)
mdel = function(x) -x^2+10*x+20
vals = mdel(concobs) + rnorm(length(concobs),0,12)
dat = data.frame(conc=concobs,val=vals)

CnrLwlss

# Testing whether small deletions are under-represented within fibres containing multiple deletion species

# Assume 200 mtDNA molecules per fibre section
Nassume = 200

# P7, P15 and P16 are the proportions of smaller mtDNA species in fibres with two or more mtDNA species
P7 = c(0.6,0.2,0.49,0.33,0.57,0.75,0.47,0.29,0.27,0.23,0.51,0.54)
N7 = rep(Nassume,length(P7))
Ndel7 = c(2,2,3,2,2,2,2,3,2,2,2,2)
P15 = c(0.39,0.73,0.37,0.17,0.43,0.53,0.54,0.57)

CnrLwlss

library(grDevices)
library(gtools)

dat = read.delim("data/RTdata.txt",sep="\t",stringsAsFactors=FALSE)
dat$PNUM = as.numeric(gsub("P","",dat$Patient))
dat$ID = sprintf("P%02d_%04d",dat$PNUM,dat$Cell.number)
dat$PAT = sprintf("P%02d",dat$PNUM)

colfunc = colorRamp(c("blue","yellow","red"),space="Lab")
colfun = function(x, alpha=1.0) {

CnrLwlss

# Read data and rename columns
dat = read.delim("mitocyto_merged_results.csv",sep=",",stringsAsFactors=FALSE)
colnames(dat) = c("value","id","channel","patient_id","patient_type")

# Specify which ids correspond to patients, and which to control
dat$patient_type = ifelse(dat$patient_id=="M1105","patient","control")
dat$patient_id = paste(toupper(substr(dat$patient_type,1,1)),dat$patient_id,sep="_")

# Specify some colours for plotting
dat$colour = "black"

CnrLwlss

mu = 5
stdev = 2
N = 10000
data = rnorm(N,mu,stdev)

pdf = function(x) dnorm(x,mu,stdev)
bestmu = function(N,x) sum(x[1:N])/N

op=par(mfrow=c(1,2))

CnrLwlss

#install.packages(c("mixOmics","RVAideMemoire"))
library(mixOmics)
library(RVAideMemoire)

# Calculate whether measure is greater in control group after scaling
# Used to colour points in VIP plots
direction=function(dt,measure){
  dts = as.data.frame(scale(dt[,-1]))
  dts$Group = dt$Group
  res = median(dts[[measure]][dts$Group=="Control"],na.rm=TRUE) > median(dts[[measure]][dts$Group!="Control"],na.rm=TRUE)

CnrLwlss

from PIL import Image
import numpy as np
import os

for fname in os.listdir("."):
    if fname.endswith(".tiff"):
        im=Image.open(fname)
        arr=np.array(im,dtype=np.uint16)
        maxval = np.max(arr[arr>0])
        minval = np.min(arr[arr>0])