Various Tips and Tricks – Training Tech Shorts

Here are some 5-minute tips and tricks we covered in today’s training.

Setting the terminal prompt using Liquid Prompt

You can enhance your terminal experience by customizing your prompt with Liquid Prompt. Liquid Prompt provides dynamic, context-aware information such as your username, active Conda environment, current directory, and Git branch directly in your prompt.

To get started, clone the Liquid Prompt repository to your home directory (or another location), then source it in your .bashrc or .zshrc. You can also enable the Powerline theme for a modern look. Here’s an example of what to add to your .zshrc:

if [[ $- == *i* ]]; then
    # Only load Liquid Prompt in interactive shells
    source ~/liquidprompt/liquidprompt
    source ~/liquidprompt/themes/powerline/powerline.theme
    lp_theme powerline
fi

Tip

If your prompt looks misaligned or displays odd characters, try switching to a Powerline-compatible font in your terminal settings.

There are many prompt customization tools available. For a comparison of popular options, see this article by the author of Liquid Prompt.

Edit Quarto Markdown tables in VSCode visual mode

The VSCode Quarto extension allows you to create and edit tables using visual mode. Visual mode provides a better interface for creating and editing tables without manually writing the table in markdown syntax.

One can switch between Source and Visual mode of Quarto extension using VS code keyboard shortcuts () or by selecting the mode from the command palette.

Adding captions to Quarto figures and tables

Here are three methods to add captions to figures and tables in Quarto.

---
title: "NBIS Report"
subtitle: "`{r} format(Sys.Date(),format='%d-%b-%Y')`"
format: html
toc: true
toc-depth: 4
number-sections: true
theme: flatly
highlight: tango
df-print: paged
code-folding: none
self-contained: true
keep-md: false
css: assets/report.css
bibliography: bibliography_library.bib
csl: nature.csl
citations-hover: true
footnotes-hover: true
crossrefs-hover: true
lightbox: true
fig-cap-location: bottom
---

```{r,include=FALSE,cache=FALSE,eval=TRUE}
## REPORT OPTIONS
## code relating to the report creation
## default working directory is the location of this document
## all code is run in the working directory as the root

# remove all variables
rm(list=ls())

# load libraries for document creation
library(knitr) # runs pandoc

# set knit options
opts_knit$set(progress=TRUE,verbose=TRUE)
opts_chunk$set(dev="svg",results="hold",fig.show="hold",fig.align="left",echo=FALSE,warning=FALSE,message=FALSE,accordion=NULL,
block.title=NULL)
#options(knitr.table.format = "html")
```

<!-- ----------------------- Do not edit above this ----------------------- -->

```{r,echo=FALSE,include=FALSE}
#species variable
my_species_name <- "*Coregonus albula*"
yourspecies_odb10 <- "actinopterygii_odb10"

```

## First way to have a caption

```{r}
#| label: tbl-buscoassembly


library(knitr)
# Create a clean formatted character vector for each line of the summary
busco_lines <- c(
  "C: 95.4% [S: 52.2%, D: 43.2%]",                 # Summary line
  "1901 Complete and single-copy BUSCOs (S)",
  "1574 Complete and duplicated BUSCOs (D)",
  "3475 Complete BUSCOs (C)",
  "35 Fragmented BUSCOs (F) (1.0%)",
  "130 Missing BUSCOs (M) (3.6%)",
  "3640 Total BUSCO groups searched (n)"
)

# Create a data frame for kable (as a single-column table)
busco_df <- data.frame(busco_lines, stringsAsFactors = FALSE)

# Render it with knitr::kable in markdown format
knitr::kable(
  busco_df,
  format = "markdown",
  col.names = NULL,
  caption = paste("BUSCO results of the assembly with the dataset", yourspecies_odb10)
)
```

## Second way to have a caption

BUSCO results :
```{.css}
table, th, td {
  text-align: left;
}
```
|     `{r} my_species_name`                                        |
|--------------------------------------------------------|
| C:88.9%[S:53.0%,D:35.9%],F:3.3%,M:7.8%,n:3640          |
| 3235 Complete BUSCOs (C)                                |
| 1928 Complete and single-copy BUSCOs (S)                |
| 1307 Complete and duplicated BUSCOs (D)                 |
| 120 Fragmented BUSCOs (F)                               |
| 285 Missing BUSCOs (M)                                  |
| 3640 Total BUSCO groups searched                        |

: Complete BUSCO results of the evidence gene build (rc1) using `{r} yourspecies_odb10` {#tbl-rc1busco}

## Third way to have a caption
```{r}
#| label: tbl-interprotableresults
#| tbl-cap: "Complete BUSCO results of the evidence gene build using `{r} yourspecies_odb10` "
#|

# Create a clean formatted character vector for each line of the summary
busco_lines <- c(
  "C: 95.4% [S: 52.2%, D: 43.2%]",                 # Summary line
  "1901 Complete and single-copy BUSCOs (S)",
  "1574 Complete and duplicated BUSCOs (D)",
  "3475 Complete BUSCOs (C)",
  "35 Fragmented BUSCOs (F) (1.0%)",
  "130 Missing BUSCOs (M) (3.6%)",
  "3640 Total BUSCO groups searched (n)"
)

# Create a data frame for kable (as a single-column table)
busco_df <- data.frame(busco_lines, stringsAsFactors = FALSE)

# Render it with knitr::kable in markdown format
kable(busco_df, format = "markdown",col.names = NULL,align = c("l", "c", "c", "c"))
```

**They are all compatible with each others**

The Here package in R

The here package in R helps you set paths. here can be used with here::i_am("$file") to set the working directory to the path where the file is hosted.

Managing R packages using PacMan

PacMan is a R package manager. PacMan is essentially a combination of the install.packages and library commands. pacman::p_load(tidyverse), for instance, sees whether you you have tidyverse installed. If you do, it loads tidyverse. If you don’t, it installs tidyverse and all dependencies, and loads tidyverse. With p_install_version you can install particular versions of the packages you are interested in. In a teaching scenario, this makes it much easier to have all students install everything that you need for your session.

The Nano editor

Nano is an editor you can use in the Terminal.

It’s loaded on Dardel using ml nano.

Setting group permissions on Dardel

Dardel doesn’t automatically set group permissions on files. So you might run into problems where group members cannot access files you have created.

Retroactively change permissions

You can update group permissions by changing file ownership to the group, and then setting the group’s permission equal to the user’s permissions:

    # thanks to Karl Johan from PDC support for this useful tidbit!
    # change the file ownership to group:
    chgrp --no-dereference --silent --recursive ${GRP} ${PWD}
    # change permission: group gets user's permissions
    chmod --silent --recursive g=u ${PWD}

This can be added to launch scripts like this run_nextflow.sh example.

Set permissions at creation

Alternatively, you can add

umask 0002

to your .bashrc, to set a file mode creation mask. Here, the last three digits encode the user, group and others classes, respectively, while the first one is ignored.

The number 0 sets read, write and execute permissions, while 2 is not granting write permission, only read and execute permission.

More information in the umask man pages and on Wikipedia.

MacOS shortcuts

Use the ditto command to copy directories while preserving permissions and timestamps.
Use history 15 to view your last 15 commands.
The banner command prints large text banners in the terminal.
Drag and drop files or folders into the terminal to paste their paths.
Use Option+Click to move the cursor anywhere in the command line.
In Finder, Cmd+I opens the info panel, where you can copy file paths or customize folder icons.

Step through code from a Pixi environment in Quarto

R packages can easily be installed in pixi environments to help resolve version incompatibilities or to ensure that others working in your project can replicate your environment.

pixi init -c bioconda -c conda-forge -p linux-64
pixi add r-base
pixi add r-tidyverse
pixi add bioconductor-phyloseq
pixi install

You can then knit your Quarto report by first activating your pixi environment containing your R installation. However, if you want to be able to view objects in your session in vs code (requires the R extension) and execute code line by line when writing your Quarto report, you can do so by including the following line in your report:

.libPaths("/path/to/pixi/directory/.pixi/envs/default/lib/R/library")

This prevents the need for activating your pixi environment before knitting the report and enables interactive coding in your Quarto report.

Using a Conda lock file in Nextflow

Managing package drift is a key challenge for reproducibility when using Conda. Typically, users create environments from an environment.yml file, sometimes pinning specific package versions. However, indirect dependencies (dependencies of dependencies) are not always pinned, which means your environment may become unresolvable or behave differently over time as upstream packages change.

To address this, it’s best practice to generate a Conda lock file, which captures the exact versions of all packages (including dependencies) in your environment. This ensures that others can recreate the environment exactly as you had it, even months later. You can generate a lock file using the --explicit flag with conda list:

# In an activated environment
conda list --explicit > spec-file.txt

Alternatively, you can use:

conda env export --name <env> --explicit > spec-file.txt

To recreate the environment from this lock file, use:

conda create --name <env> --file spec-file.txt

While this approach is helpful for reproducibility, it can become unwieldy if you need to manage many environments. Fortunately, Nextflow supports using Conda lock files (which must have a .txt extension). Simply reference your lock file in the conda directive within your process definition, and enable Conda in your Nextflow configuration:

my_process.nf

process MY_PROCESS {
  conda "${moduleDir}/spec-file.txt"

  input:
  ...
}

nextflow.config

conda.enabled = true

Place spec-file.txt in the same directory as your process script.

Tip

When building pipelines with nf-core modules, you’ll find that they often come with prebuilt containers and Conda environment files, but not with lock files. You can use Seqera Containers to build or pull a prebuilt container and extract its lock file.

For example, MultiQC has a prebuilt container. The build log shows that an environment.lock file is included inside the container.

To extract the lock file from the container we can use:

docker run --rm -it community.wave.seqera.io/library/multiqc:1.29--e3ef3b42c5f9f0da cat environment.lock > multiqc-1.29-spec-file.txt

You can now use this lock file with the conda directive in Nextflow to recreate the same environment as the Docker container.