Breakout session - Model organism
Erich Schwarz, Rocio Martinez Nunes
## Which animals are we studing?
Off target detection (CRISPRs)
Metabolomics/proteomics/mRNA (Omics integration)
UCSC genome brownser for human and mouse
With sme other animals you might
[Gbrowse](https://github.com/GMOD/GBrowse) can facilitate the visualization of genomes of more unconvencional animals once you have a reference genome.
can also be useful.And it is user friendly.
* 5' and 3' UTR are maybe not that well annotated in humans, or you can find weird problems with asserting reads position when they are located at UTRs
* Illumina kits for library preparation for RNA-seq
are kind of expensive so Rocio is also using New england kits
* RNA extraction with Trizol
* Qiagen micro RNAasy kit to clean the RNA.
## RNA-seq pair-end (PE)
For many purpose 50 SE is fine. To see general diference between two groups, for instance.
For more specif purposes, like difficult regions (UTRs, GC "rich", splice variants) PE is better.
Is it possible to do RNA-seq with pacbio?
Yes.BUt if the RNA is not high-quality, there is no pointing on trying to get the 5kb reads.
RIN: RNA integrity number, you want it to be at least 7, but higher the better.In the bioanalyzer dependending on the amount of the RNA there is different chips that you can use ( nano, pico)
## Ways to deal with batch effects
Rank genes by the level of expression with the functional
Kallisto is a piece of software that use pseudo-alignment, a super fast pseudo-mapping
sleuth is a R packages that you can model batches and can bootstraps. Which gives you a data driven assesment. Salmon can use to bootstrap to put sleuth.
bootstraping: Instead of assuming a bell curve, creatinh a random number of replicates of you samples keeping the same distribution of you sample. So, you can get an empirical idea if the signals are real.
Functional enrichment: biological meaning
### How to choose software to use ?
Easy to use
can be installed with conda
### How to manipulate data using in R
### Where look for help?
* Should not write the arguments manually, automate
* Check the version of the packages you are using it may be useful for reproducibility
* You can use bioconda to create environments with specific versions of packages/softwares