nonlinear regression and nonlinear least squares in R

a useful document

http://socserv.mcmaster.ca/jfox/Books/Companion/appendix/Appendix-Nonlinear-Regression.pdf

CV values

noise level test: bigger CV value means more noisy data
higher G (shape) parameter in model simulation reveals smaller lifespan data values.
higher R(rate) parameter in model simulation reveals higher mean lifespans.

free hard drive space command

df -H : hard drive allocation and usage check
top     : memory usage check (slow method)
vm_stat : memory usage check (faster)

Academic writing tips

Ten easy rules from Zhang et al:

http://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003453#pcbi.1003453-Watson1

today's note

1)Copy number variation. CNV is worth to study. : http://www.nature.com/scitable/topicpage/copy-number-variation-445
2)Focus C. elegans for next project as experimental lifespan data sets.
3)Data bases for lifespan :

http://kaeberleinlab.org/projects/lifespan-observations-database

*Observations database
*Sagaweb database
*Managed databases

solving ODEs

A nice recalling examples here:

http://mathinsight.org/ordinary_differential_equation_introduction

Bayesian statistics vs Maximum Likelihood estimation

Screenshot is taken from Hartig et.al 2011, Statistical inference for stochastic simulation models – theory and application

 

here is description why we are normalizing the posterior position in Bayesian statistics.

Metropolis Hastings -MCMC in R

MCMC is achieved by;

1)Starting at a random parameter value (old)

2)Choose a new parameter value which is close to the old value (current) based on some probability density and that is called future function (new)

3)hop to this new point with a probability  p(new)/p(old), where p is the target function, p>1.

https://github.com/florianhartig/LearningBayes/blob/master/CommentedCode/02-Samplers/MCMC/Convergence.md

what is noise

randomness: Living organisms flow on inferences (guesses) about the best response to make, because of the information they receive from outside world (environment) is a part of diluted noise.


                 information(about world)
                             

inferences(on present)------------------>modify behaviors (to optimize survival probability)

Integrative Genomics Viewer

http://software.broadinstitute.org/software/igv/download

IGV is an excellent way to visualize seq. data , whether it is whole genome seq or ChIP-seq or RNA-seq.

data analysis course sequences on edX

https://courses.edx.org/courses/course-v1:HarvardX+PH525.7x+3T2015/cd8cfac0f386436fa0cb1ed3d0012328/

GitHub commands on MAC OSX

Nice link on when and how to perform statistical tests:

What statistical analysis should I use?

http://www.ats.ucla.edu/stat/stata/whatstat/whatstat.htm