DenverFINRISKHacky

Hackathon for the DREAM FINRISK Challenge cardiovascular event prediction from microbiome

Placeholder README for the Hackathon taking place in Denver 14th to 18th November 2022.

Things to consider:

Installed software on your computer (preferably laptop):

R (4.2.1 or preferably newer)
R Studio suggested if you want a bit of assistance
Highly recommend git, so we can merge and share code easily
GitHub (where you're currently at) will serve as our main platform

R packages & functions

Installation from CRAN or BioConductor

Install packages from the terminal with install.packages("pckgName") (replace pckgName with the exact name of the package you want to install. Note that this only works for Central R Archive Network packages (CRAN).

For Bioconductor you need to use:

if (!require("BiocManager", quietly = TRUE))
    install.packages("BiocManager")

BiocManager::install("ComplexHeatmap")

Recommended R packages

Packages (and for what purpose they could be used):

ggplot2 (for visualizations beyond base visualizations)
ComplexHeatmap (for advanced heatmap visualizations)
glmnet (for regularized regression with L1/L2 norms)
...

R functions to check out

name of R function; description of the purpose (if it needs a non-base package, package name indicated here)

write.table; for writing output
read.table; for reading output
setwd; for setting current working directory, notice that R understands this as your root directory
|>; R's native pipe operator (R >= 4.3.0)
plot; an ordinary x-axis vs. y-axis plotting
boxplot; a boxplot
ls; list objects in current R workspace
<-; placement operator (recommended for using when not talking about function parameters)
=; placement operator or function parameters (recommended to use only with function parameters)
...

Advanced stuff to consider

Learning how to use Singularity containers: https://docs.sylabs.io/guides/3.5/user-guide/introduction.html (the DREAM Challenge comes with an example container)

Literature regarding the topic

For example, reading

Salosensaari A, Laitinen V, Havulinna AS et al., Taxonomic signatures of cause-specific mortality risk in human gut microbiome. Nat Commun 12, 2671 (2021). https://doi.org/10.1038/s41467-021-22962-y would be useful.

Synapse and the DREAM

Registration on Synapse platform is highly encourated:

Overview to the Challenge phases https://www.synapse.org/#!Synapse:syn27130803/wiki/619280

Data description is essential for understanding what we're trying to model: https://www.synapse.org/#!Synapse:syn27130803/wiki/619274

Objects in train_test.RData

train_x / test_x : rank inverse normal distributed counts
train_raw / test_raw : read counts before above normalization
train_p / test_p : phenotype (clinical) info for train and test sets respectively
train_y / test_y : Surv-response vector for train and test sets respectively

Short citations / insight

"Valtimokovettuma tautiin eli ateroskleroosiin ja siihen liittyviin sydän- ja verisuoniongelmiin on mikrobiomilla tutkittuja yhteyksiä. Asteikolla 1-10 mitattuna, mitä enemmän suojaavia tekijöitä löydetään mikrobiomista, sitä korkeampi on suojaava vaikutus näihin sairauksiin. Jotkut bakteerit, kuten Erysipelotrichaceae tuottaa trimetyyliamiinia (TMA) koliinista, jota löytyy kananmunan keltuaisesta, sekä L-kartiniinia, jota saadaan punaisesta lihasta ja kalasta. TMA hapetetaan maksassa trimetyyliamiini-typpioksidiksi TMAO, johon liittyy kohonnut riski sairastua ateroskleroosiin ja muihin sydän- ja verisuonitauteihin." (https://www.riskigeeni.net/mikrobiomi.html)

-> Some bacteria such as Erysipelotrichaeae produce trimethylaminine (TMA) from coline and L-cartinine; TMA is oxidated in the liver to TMAO, which involves a heightened risk to suffer from ateroscloreosis and other cardiovascular diseases

prashiyn/finrisk_hackathon