|
|
## Workflow
|
|
|
|
|
|
![](https://ascgitlab.helmholtz-muenchen.de/chromatin_summer_school_2022/workflows/snakemake-ngs-processing/-/raw/main/images/rulegraph.png)
|
|
|
|
|
|
- Login to the HPC
|
|
|
```{bash}
|
|
|
ssh your.user@hpc-submit01.scidom.de
|
|
|
```
|
|
|
|
|
|
- Change to your workflow directory
|
|
|
```{bash}
|
|
|
cd /lustre/groups/shared/chromatin/user.name/snakemake-ngs-processing/
|
|
|
|
|
|
# if your run did not finish
|
|
|
# cd /lustre/groups/shared/chromatin/gabriela/workflows/snakemake-ngs-processing/
|
|
|
```
|
|
|
|
|
|
- Check sub-folders
|
|
|
```{bash}
|
|
|
ls -l
|
|
|
ls -l results
|
|
|
```
|
|
|
|
|
|
![Screenshot_2022-08-26_at_09.09.50](uploads/1231eda788685aa5117dc639b7945b1e/Screenshot_2022-08-26_at_09.09.50.png)
|
|
|
|
|
|
- Browse result files
|
|
|
```{bash}
|
|
|
ls -l results/mapped_reads/hg38/*
|
|
|
ls -l results/peak_calling/hg38/macs2/narrow/*
|
|
|
```
|
|
|
- Find log files
|
|
|
```{bash}
|
|
|
ls -l results/logs
|
|
|
```
|
|
|
|
|
|
- Where do you find the `bowtie2` log files?
|
|
|
- How many reads are aligned concordantly exactly 1 time in the `CT_H3K27ac_4` file?
|
|
|
- What is the median `fragmentSize` of `Gr5_input` sample?
|
|
|
|
|
|
```{bash}
|
|
|
ls -l results/fragmentSize/hg38/
|
|
|
```
|
|
|
- Download `png` files to your local computer (via `vscode` or `scp`)
|
|
|
- compare fragment sizes to the `BioAnalyzer`results (files are in Slack)
|
|
|
- compare Cut&Tag fragment sizes of H3K27me3 and H3K27ac
|
|
|
|
|
|
![image](uploads/08501eb0e5408204c9450d35d7a7280f/image.png)
|
|
|
|
|
|
|
|
|
## MultiQC Report
|
|
|
|
|
|
```{bash}
|
|
|
ls -l results/qc/
|
|
|
```
|
|
|
|
|
|
- Download the `multiqc` folder to your local computer (via `vscode` or `scp`)
|
|
|
- Open the `multiqc_report.html`
|
|
|
|
|
|
![Screenshot_2022-08-26_at_09.25.26](uploads/b652cc951b393475048701d2ed491a64/Screenshot_2022-08-26_at_09.25.26.png)
|
|
|
|
|
|
- Which sample has the most and the least number of sequenced reads?
|
|
|
- Which sample has the most uniquely mapped reads?
|
|
|
- Which assay contains more PCR duplicates?
|
|
|
- Among Cut&Run which H3K27me3 sample/group has the least PCR duplicates?
|
|
|
- Which sample has the highest adapter content (look up its id in the `sample-manifest.tsv` file)?
|
|
|
- Why is it "bad" to have so many adapters if we remove them anyway?
|
|
|
|
|
|
|
|
|
## IGV session
|
|
|
|
|
|
- Open `IGV` browser
|
|
|
- Load the `bigwig` file (downloaded yesterday) with the most and the least PCR duplicates among Cut&Tag samples
|
|
|
- Where do you find these samples on the correlation [heatmap](https://ascgitlab.helmholtz-muenchen.de/chromatin_summer_school_2022/workshops/correlation/-/blob/main/plots/heatmap_SpearmanCorr_readCounts.png)?
|
|
|
- select samples with different fragment sizes of the same assay
|
|
|
- open their `bigwig` files in ´IGV´
|
|
|
- check whether they cluster on the correlation heatmap
|
|
|
- Look at IgG samples that have low correlation
|
|
|
|
|
|
---
|
|
|
|
|
|
- coverage correlation
|
|
|
![heatmap_SpearmanCorr_readCounts](uploads/2ec90f322b23cd350a3d3b0d91d3ea57/heatmap_SpearmanCorr_readCounts.png)
|
|
|
|
|
|
- peak Jaccard
|
|
|
![Unknown](uploads/773af2fb31e27e52cb27a7ebc9138311/Unknown.png)
|
|
|
|
|
|
|