This tutorial & exercise is designed for trainees interested in joining our lab. It reflects our expected computing skills using R, Python, Linux shell commands and bioinformatics workflow languages. Even if you lack the skills for one or more of the languages at the time you start this tutorial, we believe the learning curve for new skills enough to complete the exercises is reasonable given a few hours of effort. Still, please do not hesitate to contact us (wang.gao@columbia.edu) if there is a blocker as you go through the material.
An additional note on effort: for student interns and research assistants we expect a minimum of 10hrs/week effort. Successful completion of this exercise in a week's time is an indication of the efforts you are able to devote.
In this task you are going to work with git from command shell, and install basic software and packages needed for data analysis of Tasks 2 and 3.
Most of our work will be saved and shared on github in public or private repositories. If you have not used git in the past, please follow the instructions here for a 5 minutes git tutorial.
As the next step please fork this repository, add your name to the Markdown file named hello.md, commit it to github with a customized commit message, eg, "Add my name and github handle", and create a pull request so we can see your update and incorporate it to the repository.
This tutorial (and our research in general) requires R, Python, Script of Scripts (SoS) bioinformatics workflow system and docker.
Please follow this setup instruction to complete the installations.
This task is an example of a bioinformatics workflow developed at our group. It uses IPython notebook (with JupyterLab IDE as a recommendation), and runs an SoS kernel for bioinformatics workflows.
Please find the example notebook file notebook/orientation.ipynb, follow the instructions and complete the Quiz at the end of the notebook.
Please follow the instructions and complete the R exercise analysis/orientation.Rmd. Rmd stands for R Markdown. They are text file with R code and narratives that you can open and analyze using software such as Rstudio.
Please follow the instructions and complete the genetic fine-mapping exercise notebook/finemapping.ipynb. This is somewhat advanced material for those who has a background in statistical genetics.
Organizing and communicating your work with others is essential to your success in conducting reproducible computational research. For every project we require the analysis written in Rmd and IPython notebooks to be converted to a research website that will either be hosted on github (a service github provides) or accessed locally on one's web browser.
We will use workflowr to organize the R analysis. Please follow workflowr instructions to convert the Rmd files under analysis/ folder into HTML based website under a directory called docs.
You should find a file docs/index.html and view it in your web browser once you successfully build the website.
We will use jnbinder program to organize the Python analysis. Please follow the section "Run from a docker image" up to step 3 and return here, then type:
jnbinder --root docs/ipynb
The first time you run the command above, a website template will be configured under docs/ipynb and the program will quit on error with a line of prompt instructing you to edit
a file called config.yml. Please open this file with text editor and uncomment to configure name, repo and footer as you see fit, and set add_commit_info to False for now.
Finally for include_dir please specify ["notebook", "workflow"] to include notebooks under these folders to the website to be generated.
Now please run the command above again to generate IPython based HTML files to docs/ipynb folder. You can view docs/ipynb/index.html in your web browser, particularly please click and view the tabs "Notebook" and "Workflow".
After you have completed all tasks please make a tarball for your docs folder and email it to wang.gao@columbia.edu for us to review.
Upon submission of your work, we will review and determine a fit for you based on your response. Projects available for Spring 2021 include:
Methods with applications:
- Rare variant association analysis in family data
- Method for gene-gene / gene-environment interaction in large-scale association studies
- Inference of kinship in ancient DNA data
- Statistical fine-mapping in family-based study designs
- Statistical fine-mapping in multiple ancestries
- Multi-tissue alternative polyadenylation calling from brain RNA-seq data
- Statistical benchmark for colocalization analysis
- Multivariate models for colocalization analysis
- Copy number variation association study for Alzheimer's disease
- Evaluation of fine-mapping in unbalanced case-control design
- Matrix completion and prediction of Alzheimer's disease risk using epidemiological records
- Alzheimer's disease risk prediction and subtyping using whole genome sequence data
Applied data analysis:
- Rare variant association studies with exome data in Alzheimer's disease families
- Linkage analysis for Alzheimer's disease families
- Gene-gene and gene-environment interaction analysis in UK Biobank data for Asthma
- Pleiotropy and mediation analysis in UK Biobank data
- Development of polygenic risk score prediction pipeline
- Statistical fine-mapping with functional genomic annotations
- Linear mixed model analysis of various phenotypes in UK Biobank data
- Improve and implement fine-mapping pipeline for UK Biobank data analysis (univariate and multivariate)
- Multi-tissue fine-mapping for splice QTL data in GTEx project
- Pipeline for TWAS analysis
- ASoC variants and DNase footprint calling in ROSMAP data
Software development:
Please also find here a list of past and ongoing projects in our group.