1 Session 1: Initial diagnostic and essential commands

1.1 Overview

Biological question

What information can we recover from an unfamiliar sequence file?

This first session has two jobs. It gives everyone a shared mental model of how a computer works and what a compute cluster is, and it runs a short diagnostic so the rest of the course can be adjusted to the group. No prior experience is assumed. If you have used a terminal before, help the person next to you.

Starting-point survey (about 4 min)

Before the lecture, fill in the short survey: starting-point survey, or scan the code below.

It asks about your background, the tools you have already met, and the data you care about. Your name is optional and the survey is not graded: the answers are used to weight the examples for the rest of the course, and the class summary is shown back to you a few minutes later so you can see how mixed the starting levels are.

This is separate from Section 1 of your worksheet, a self-assessment you fill in at the end of the practical (block G). Nothing in Sessions 1 and 2 is graded.

1.1.1 Learning outcomes

By the end of this session, you should be able to:

name a few common types of biological data and say roughly how each is produced;
name the main parts of a computer and say what each one does;
explain what a compute cluster is and why we use one;
connect to the university cluster through RStudio Server (and know that ssh exists);
distinguish the RStudio Console from the Terminal;
identify your current folder and list its contents;
create and enter a folder;
copy a file and look inside it with head, tail, grep, and wc;
recognise FASTA records, count them reliably, and spot one that is not DNA or was truncated;
read an error message and correct the command;
keep a command log that another person can rerun from a clean start.

1.1.2 Session plan

Time	Block	What happens
14:15-14:20	Welcome	Anonymous starting-point survey (Form).
14:20-14:32	Theory 1	What bioinformatics is, and where biological data comes from.
14:32-14:45	Theory 2	How a computer works (live teardown).
14:45-15:00	Theory 3-4	From your laptop to a cluster. The FASTA format. Practical briefing.
15:00-15:15	Practical A	Connect to RStudio Server. Console vs Terminal. Start a command log.
15:15-15:35	Practical B	Navigate folders: predict each line, then run it.
15:35-15:50	Practical C	Copy the file. Four ways to look inside it.
15:50-16:00	Checkpoint	Save your work. Short break.
16:00-16:22	Practical D	The investigation: what is wrong with `mystery.fasta`? (pairs)
16:22-16:40	Practical E	Break it and diagnose: six failures, six causes. (pairs)
16:40-16:50	Practical F	Reproduce a partner’s script from an empty folder.
16:50-17:00	Practical G	Individual wrap-up (not graded).

1.2 Theory (≈45 min)

1.2.1 1. Bioinformatics and the data it works on

Bioinformatics connects a biological question to data and to a reproducible analysis that produces a result you can explain.

Code

flowchart LR
  Q["Biological question<br/>(e.g. which genes respond to stress in this organism?)"] --> D["Data<br/>(files: FASTA, FASTQ, tables…)"]
  D --> A["Analysis<br/>(a few commands, run in order)"]
  A --> R["Result<br/>(a number, a table, a figure)"]
  R --> I["Interpretation<br/>(answer + one limitation)"]

flowchart LR
  Q["Biological question<br/>(e.g. which genes respond to stress in this organism?)"] --> D["Data<br/>(files: FASTA, FASTQ, tables…)"]
  D --> A["Analysis<br/>(a few commands, run in order)"]
  A --> R["Result<br/>(a number, a table, a figure)"]
  R --> I["Interpretation<br/>(answer + one limitation)"]

The analysis is just a list of commands run in a defined order. If you write the commands down, another person (including future you) can run them again and get the same result. That is what “reproducible” means, and it is the habit this course builds.

1.2.1.1 Where does biological data come from?

Different instruments answer different questions. You do not need to master any of these today: the goal is to recognise the main families and the kind of file each one lands in.

Data type	How it is produced	Typical files	Example question	In this course
Genome / amplicon sequencing	High-throughput sequencers (Illumina, Oxford Nanopore) read DNA fragments	FASTA, FASTQ	Which species are in this soil sample?	Sessions 2-3, 5-6
Bulk RNA-seq	Sequencing of RNA pooled from many cells → one expression value per gene	FASTQ → count table (CSV/TSV)	Which genes respond to drought?	Session 8 (project)
Single-cell RNA-seq	RNA sequenced per individual cell	sparse matrix (MTX, H5)	Which cell types are present, and how do they differ?	mentioned only
Proteomics	Mass spectrometry of peptides	mzML → abundance table	Which proteins change with treatment?	mentioned only
Metabolomics	Mass spectrometry or NMR of small molecules	feature-intensity tables	Which metabolites mark a phenotype?	mentioned only
Imaging / microscopy	Automated microscopes + image segmentation	TIFF / OME-TIFF → measurement tables	Does cell shape change between conditions?	mentioned only
Genotyping / GWAS	SNP arrays or sequencing → genotype matrix	VCF, genotype tables	Which variants associate with a trait?	Session 6 / project

The pattern to remember: whatever the instrument, the analysis almost always ends up in one of two shapes of text file: sequences (FASTA, FASTQ) or tables (rows = samples or features, columns = variables). This course spends its time teaching you to open, check, and summarise those two shapes with confidence.

Bulk vs single-cell, in one line each. Bulk measures the average over thousands of cells: cheap, and enough to ask “which genes differ between conditions”. Single-cell resolves cell-to-cell variation and rare populations: more expensive, and the data are much sparser. How the sequencing itself works (library preparation, paired-end reads, read length) comes in Session 2, alongside FASTQ and quality control.

1.2.2 2. How a computer works

A computer has a small number of parts. Each one has a single job.

Part	Job	Kitchen analogy
CPU (processor)	Does the calculations, one step after another (very fast).	The cook
RAM (memory)	Holds the data currently being worked on. Fast, but erased when the power goes off.	The worktop: only what you are using right now fits on it
Disk / SSD (storage)	Keeps files when the power is off. Slower than RAM, much bigger.	The fridge and cupboards
GPU (graphics card)	Does many simple calculations in parallel. Used for images and for deep learning.	A brigade of line cooks all chopping at once
Operating system (Linux, macOS, Windows)	Manages the parts, the files, and the programs.	The kitchen manager
Network card	Talks to other computers.	The delivery door

Two ideas matter for the rest of the course:

A program is just a file that the CPU knows how to run. ls, head, R and blast are all programs sitting on the disk.
Files live in folders, folders live in folders. The chain of folder names that leads to a file is its path, for example /home/you/bioinfo_course/data/reads.fasta. The leading / is the top of the tree. The folder you are “standing in” right now is the working directory; a command that is given no path looks there.

1.2.2.1 The shell

The shell lets you drive the computer by typing commands instead of clicking. You type the name of a program, press Enter, and read what it prints back. If you know R, this is the same idea with a different language. The shell we use is called bash. It is good at three things bioinformatics needs constantly: navigating files, repeating an analysis exactly, and moving or summarising large files automatically.

1.2.3 3. From your laptop to a compute cluster

A laptop is fine for writing code and looking at results. It is not fine for aligning a million sequencing reads: not enough CPU, not enough memory, and the data would not even fit. A compute cluster solves this. It is a set of powerful computers (“nodes”) in a server room that many people share, with the large data stored right next to them.

For this course we use LEGcompute, hosted by the Laboratory of Evolutionary Genetics and shared across the Institute of Biology (crolllab/IBIOL-compute-resources).

Resource	LEGcompute
CPU cores	~1000 (Intel + AMD)
GPUs	RTX 3090, 3× RTX A5500, 4× RTX 5000 Ada, 2× RTX PRO 6000 (96 GB)
Memory	4 TB RAM
Storage	80 TB SSD
System	Ubuntu 26.04 LTS, Slurm scheduler, RStudio Server

1.2.3.1 How the cluster is organised

Code

flowchart LR
  L["Your laptop"] -- "https :8787 (RStudio)" --> H
  L -- "ssh" --> H
  H["Login / head node<br/>(where you land)"] -- "sbatch / srun" --> S["Slurm scheduler"]
  S --> C1["compute node"]
  S --> C2["compute node"]
  S --> G["GPU node"]
  H --- ST[("Shared storage<br/>/home  /data  /scratch")]
  C1 --- ST
  C2 --- ST
  G --- ST

flowchart LR
  L["Your laptop"] -- "https :8787 (RStudio)" --> H
  L -- "ssh" --> H
  H["Login / head node<br/>(where you land)"] -- "sbatch / srun" --> S["Slurm scheduler"]
  S --> C1["compute node"]
  S --> C2["compute node"]
  S --> G["GPU node"]
  H --- ST[("Shared storage<br/>/home  /data  /scratch")]
  C1 --- ST
  C2 --- ST
  G --- ST

You log in to the head node. Everyone shares it, so it is for light interactive work only: editing, plotting, small commands (everything we do today).
Real jobs are submitted to Slurm with sbatch, which sends them to a free compute node so the machine is never overloaded and everyone gets a fair share. We meet Slurm properly in a later session.
Storage is shared by every node, so a file you create on the head node is visible in your jobs too:
- /home/username: scripts and small files (quota: 100 GB per student);
- /data/username: large datasets and analysis output (1+ TB);
- /scratch: temporary space; files older than 30 days are deleted automatically.
~200 bioinformatics tools are pre-installed as modules (module load SAMtools). Nothing to install during class.

No backups

No backup is made of anything on the server. Keeping a copy of work you care about is your responsibility.

1.2.3.2 Access for this course

Connecting to LEGcompute

1. Be on the university network. Either the unine WiFi on campus, or the GlobalProtect VPN from home. There is no way in from the open internet.

2. Claim your account. Open the shared Google Sheet the instructor. On the course tab, pick a free username and write your name in the yellow cell next to it. The first page of the sheet explains how to set or reset your password (via the user account server).

3. Connect (pick one):

Method	How	Good for
RStudio Server	Browser → http://legcompute.unine.ch:8787 → log in. Use the Terminal tab for the shell.	Everyone today. Nothing to install.
SSH	In a local terminal: `ssh username@legcompute.unine.ch`	Later, once you are comfortable.

We use RStudio Server for the whole course. It gives you an editor, an R Console, a Terminal and a file browser in one browser tab, with nothing to install.

1.2.4 4. The FASTA format

FASTA is the simplest way to store one or more biological sequences as text. A record has two parts:

a header line, whose first character is >, giving the sequence a name;
one or more sequence lines (the bases or amino acids).

>NC_045512.2 Severe acute respiratory syndrome coronavirus 2, surface glycoprotein region (partial)
ATGTTTGTTTTTCTTGTTTTATTGCCACTAGTCTCTAGTCAGTGTGTTAATCTTACAACC
AGAACTCAATTACCCCCTGCATACACTAATTCTTTCACACGTGGTGTTTATTACCCTGAC
>oligo_F short PCR primer
GATTACAGGCGTGAGCCACCACGT

Here the first record’s sequence is wrapped over two lines; the second fits on one. Both are still single records.

The next record starts at the next line beginning with >. Counting records is therefore the same as counting > lines. Nothing in the format tells you whether a sequence is DNA or protein, how long it is, or whether it is complete: you have to look.

1.2.5 Practical briefing

You will connect to the cluster, build a small project folder, and copy in one FASTA file, mystery.fasta. It is deliberately messy: different line widths, a record that is not DNA, and one that was cut off mid-download. Your job is to find that out with the commands, in pairs. Write every command you run into a script file, even (especially) the ones that fail: an error message is evidence, not a reason to start over.

1.3 Practical (≈1 h 50)

1.3.1 Essential commands for today

Six commands to move around, plus four to look inside a file. That is the whole toolbox for today.

Command	What it does	Example
`pwd`	print working directory: “where am I?”	`pwd`
`ls`	list the current folder (`ls -lh` adds sizes)	`ls -lh`
`mkdir`	make a new folder	`mkdir data`
`cd`	change directory: move into a folder	`cd data`
`cp`	copy a file: `cp <what> <where>`	`cp reads.fasta .`
`head`	show the first lines of a text file	`head -n 4 reads.fasta`
`tail`	show the last lines	`tail -n 3 reads.fasta`
`wc -l`	count lines	`wc -l reads.fasta`
`grep "^>"`	show the lines that start with `>` (FASTA headers)	`grep "^>" reads.fasta`
`grep -c "^>"`	count the lines that start with `>` (records)	`grep -c "^>" reads.fasta`

^> means “a line beginning with >”. cat (whole file) and less (scroll, press q to quit) are useful too. Deleting files is deliberately not covered today.

Three habits that save time

Tab completion. Type the first few letters of a command or file name and press Tab. The shell finishes it for you and prevents typos.
Command history. Press ↑ to bring back the previous command and edit it.
Send a line from the script. With your .sh script open, put the cursor on a line and press Ctrl+Enter (Cmd+Enter on macOS) to run it in the Terminal without copy-pasting.

1.3.2 A. Connect and get oriented (15:00-15:15)

In a browser, go to http://legcompute.unine.ch:8787 and log in with your cluster username and password.

The RStudio Server login page.
Find these four parts of the RStudio window:
- Source editor (top left): where scripts are written;
- Console (bottom left): runs R code;
- Terminal (a tab next to the Console): runs shell commands;
- Files pane (bottom right): a click-through file browser.
If there is no Terminal tab, open one with Tools → Terminal → New Terminal. The welcome text it prints is just a message from the server.

Console vs Terminal

Everything in this practical is typed in the Terminal. The Console is for R and comes in later sessions. If a shell command gives a strange R error, you are in the wrong panel.

Open a new script: File → New File → Shell Script. Save it straight away as session_01_commands.sh. From now on, every command you try goes in this file, including ones that fail. Keep it: you reuse it in Session 2, and a partner reruns it in block F.

If you cannot connect

Do not lose ten minutes alone. Pair up with a neighbour and share one screen, or use a room computer, and tell the instructor. Fixing access for one person should not stop the class.

1.3.3 B. Where am I? Navigating folders (15:15-15:35)

Work one line at a time. Before you press Enter, write in your script, as a # comment, what you expect to see. Then run it and compare. A wrong prediction is the useful part: it shows exactly which idea to fix.

pwd                     # predict: which folder does this print?
ls                      # predict: what is already here?
mkdir bioinfo_course    # predict: what does ls show now?
cd bioinfo_course
pwd                     # predict: how has the path changed?
mkdir session_01
cd session_01
pwd

Checkpoint

pwd ends in bioinfo_course/session_01.

Make a data folder and move into it. Predict each line first.

mkdir data
cd data
pwd

Checkpoint

pwd ends in bioinfo_course/session_01/data. Check this before block C: the cp commands below copy into whatever folder you are currently standing in, so if pwd shows the wrong place, cd there first, or the files land somewhere you will not find them.

1.3.4 C. Get the data and take a first look (15:35-15:50)

Make sure pwd still ends in bioinfo_course/session_01/data (you are copying files into this folder, not into the course materials folder), then copy the sequence file and your worksheet, and check they arrived:

cp /home/cmaslard/bioinfo_tools_course/session_01/mystery.fasta .
cp /home/cmaslard/bioinfo_tools_course/session_01/WORKSHEET.md .
ls -lh

The final . means “here, in the current folder”. ls -lh also shows the file sizes.

WORKSHEET.md is a plain text file of questions, not a script: open it from the Files pane (click its name) or File → Open File, and type your answers directly under each “Answer:” line as you go through today’s blocks. Save it like any other file (Ctrl/Cmd+S). There is nothing to submit: it stays in your folder, and your instructor may look at it to see where the class is.

Now look inside it four ways. Predict the output of each line before you run it.

head -n 6 mystery.fasta       # the first few lines
tail -n 3 mystery.fasta       # the last few lines
grep -c "^>" mystery.fasta    # how many records (lines starting with >)
wc -l mystery.fasta           # how many lines in total

Checkpoint

grep -c "^>" prints 40. wc -l prints 134. Keep both numbers: block D asks why they differ. Note the output of pwd and the ls -lh line for mystery.fasta now, for Section 2 of your worksheet.

Moving files in and out (no exercise, just so you know)

You will not need it today, but the RStudio Files pane moves files between your laptop and the server: Upload (laptop to server) and tick a file then More -> Export… (server to laptop). No scp required.

Uploading a file through the RStudio Files pane.

1.3.5 15:50-16:00: Checkpoint and short break

Before you step away: your script is saved, and pwd ends in bioinfo_course/session_01/data.

1.3.6 D. The investigation, in pairs (16:00-16:22)

mystery.fasta holds 40 records. Most are unremarkable filler; seven are not, and they are scattered in among the rest, not conveniently grouped at the top. Work with one other student. For each question below, write the answer and the command you used in Section 3 of your worksheet.

How many sequence records are in the file? Which command answers this without miscounting, and why is scrolling and counting > by eye unreliable here?
wc -l gave 134, not 80 (which would be two lines per record). Give the two separate reasons.
One record is not DNA. Which one? Give two clues, one from the sequence and one from the header.
One record was cut off during download. Which one, and how does tail show it?
Look at contig_07. Its sequence mixes UPPERCASE and lowercase and contains a run of N. What do the lowercase letters mean, and what does the N run mean? You do not need to change anything.
Which record has the shortest complete sequence? (The cut-off record from question 4 does not count: it has no sequence at all.) Show it and say how you decided. There is no single command that measures sequence length for you here: scroll through with cat or less, spot the records whose sequence fits on one line, and compare those by eye.

Short on time?

At 16:22, move to block E with whatever you have, even if some questions are unanswered or your answers are unsure. Partial answers are fine: this is not graded, and there is no benefit in rushing block E to finish block D.

Commands you may want

grep "^>" mystery.fasta          # list every header at once
grep -c "^>" mystery.fasta       # count records
head -n 20 mystery.fasta         # a first chunk of records, to get a feel for the file
tail -n 3 mystery.fasta          # last lines
grep -A 1 "SPIKE" mystery.fasta  # a matching header and the line right after it

1.3.7 E. Break it and diagnose, in pairs (16:22-16:40)

Every command below fails or surprises you, each for a different reason. For each one: predict what will happen, run it, copy the exact message, then write a version that works. Record all of it in your script.

cd Data
head mystery.fna
head -n mystery.fasta
cp mystery.fasta
cd mystery.fasta
head ../mystery.fasta

Hints:

Linux distinguishes uppercase from lowercase.
A file name must match exactly, extension included.
Some options need a value; some commands need more than one argument.
.. is “the folder above”; a bare name is “in the current folder”.

Return to bioinfo_course/session_01/data when you are done.

Short on time?

At 16:40, move to block F even if you have not diagnosed all six. Whatever you have recorded, including the ones you did not solve, is still useful material for block G.

1.3.8 F. Reproduce it (16:40-16:50)

Trade session_01_commands.sh with your partner. From your home folder, start clean and run their commands in order:

cd ~
mkdir replay
cd replay
# now run your partner's commands, from the top, one by one

Does it run start to finish? If it stops, the missing step is something they did but did not write down, usually a cd or the cp path. Add the missing lines to your own script so it works from an empty folder. That is what “reproducible” means in practice.

1.3.9 G. Individual wrap-up (16:50-17:00)

On your own, complete Sections 1, 4, and 5 of your worksheet:

one command that failed today: the exact message, the cause, and the fix;
two or three sentences: from mystery.fasta alone, what can you say about these sequences, and what can you not say?

You should leave with:

WORKSHEET.md with your answers;
session_01_commands.sh with every command you attempted, errors included, that runs from a clean start.

Session 1 is a diagnostic, not a test. It is not graded: the continuous assessment (a short Moodle quiz at the end of each session) starts at Session 3. Your instructor may still look at your worksheet to see where the class is.

Keep the bioinfo_course folder. Session 2 reuses it.

1.4 Checkpoint

You are done when all of the following are true:

you reached the Terminal in RStudio Server and know it is not the Console;
session_01_commands.sh is saved, includes the commands that failed, and runs from an empty folder;
pwd ends in bioinfo_course/session_01/data, and that folder contains mystery.fasta;
you can say the file has 40 records, that grep -c "^>" is the reliable way to count them, and why wc -l gives a different number;
you can name the record that is not DNA and the record that was truncated, and say how you spotted each;
you can explain in one sentence that a FASTA record starts on a line whose first character is >.

1.5 Optional challenge

For your own interest, not graded.

A. Look at a real genome without opening it. LEGcompute keeps a shared folder of reference genomes that every account can read, no path from the instructor needed:

ls -lh /data/genomes/fasta/
head /data/genomes/fasta/Vitis_vinifera.12Xv2.dna.toplevel.fa
grep -c "^>" /data/genomes/fasta/Vitis_vinifera.12Xv2.dna.toplevel.fa

Note the file size, how many sequences grep -c finds, and how fast head answers even on a file of hundreds of megabytes. Why would opening it in the editor be a bad idea?

B. Make the answers reproducible. Write a three-line script answers.sh that prints the record count and the last record of mystery.fasta, then run it with bash answers.sh.

1.6 Going further: more shell commands

Optional, if you have time left

Not part of the essential route, not graded, and nothing later in the course assumes you have done it. If you reach 16:50 and blocks A to G are done, work through as much of this as you like; stopping partway through is completely fine.

A few more small, useful commands. Try each one in bioinfo_course/session_01/data.

1.6.1 More ways to `ls`

ls -l     # long format: permissions, owner, size, date
ls -a     # include hidden files (names starting with a dot)
ls -h     # human-readable sizes, e.g. 6.4K instead of 6460
ls -lh    # options can be combined
ls -R     # list this folder and every folder inside it

Run each one and compare. Which combination would you reach for to check a file’s exact size before copying it somewhere with a tight quota?

1.6.2 Writing and reading small text files

echo prints text. On its own it writes to the screen; with > it writes to a file instead, replacing whatever was there. With >> it adds a line without erasing the rest.

echo "first line" > notes.txt
cat notes.txt              # cat shows the whole file at once
echo "second line" >> notes.txt
cat notes.txt
head -n 1 notes.txt        # compare with cat: only the first line

> and >> are two characters apart and easy to mix up: > replacing a file you meant to add to is a classic way to lose work, so pause before you press Enter on one.

1.6.3 Moving and renaming with `mv`

cp leaves the original in place; mv does not; it moves or renames.

mkdir backup
cp mystery.fasta backup/
ls backup/
mv backup/mystery.fasta backup/mystery_copy.fasta   # mv also renames
ls backup/

1.6.4 Two more worth knowing exist

less mystery.fasta scrolls a file a screen at a time with the arrow keys; press q to quit. Handy for a file too long for head to feel like enough.
nano notes.txt opens a small in-terminal text editor: type, Ctrl+X to exit, y to save. You already have an editor in RStudio’s Source pane, so you may never need it here, but you will meet nano again on other servers.

1.7 Answers

Open this after you have tried the exercises yourself

Session 1 is not graded, so there is no reason to hide the answers once you have genuinely tried each block. Use this to check your own reasoning, not instead of it: the value is in working out why, not in reading the line below.

Block C checkpoint. grep -c "^>" mystery.fasta prints 40; wc -l mystery.fasta prints 134.

Block D, the investigation.

40 records, from grep -c "^>" mystery.fasta. Counting > by eye is unreliable here because 40 headers are spread over 134 unevenly wrapped lines: easy to double-count or skip one while scrolling.
wc -l gives 134, not 80 (2 x 40), for two reasons: (a) most sequences wrap over two or more lines, adding lines beyond one per record; (b) truncated_record contributes only its header line, with no sequence line at all.
Not DNA: sp|P0DTC2|SPIKE_SARS2. Sequence clue: it contains letters that are not A/C/G/T/N (E, F, I, K, L, M, P, Q, R, S, V, W, Y). Header clue: sp|...| is a UniProt (Swiss-Prot) protein identifier, and the description says “glycoprotein”.
Truncated: truncated_record. tail -n 3 mystery.fasta shows the file ending on a header line, with no sequence line after it.
In contig_07, lowercase letters mark soft-masked regions (repetitive sequence flagged by a masking tool, not removed). The run of N is an assembly gap: the length is estimated but the bases are unknown.
Shortest complete sequence: oligo_F (24 bases, one line). truncated_record has no sequence at all, so it counts as “cut off” (question 4), not “shortest”.

Block E, six failures.

Command	Fails because	A version that works
`cd Data`	case sensitivity: the folder is `data`	`cd data`
`head mystery.fna`	wrong name: the file is `mystery.fasta`	`head mystery.fasta`
`head -n mystery.fasta`	`-n` needs a number; it consumed `mystery.fasta` as the count	`head -n 5 mystery.fasta`
`cp mystery.fasta`	`cp` needs a destination as well as a source	`cp mystery.fasta backup.fasta`
`cd mystery.fasta`	`mystery.fasta` is a file, not a directory	`cd .` (stay), or `cd ..` to go up
`head ../mystery.fasta` (from `data/`)	`..` is `session_01/`, where the file is not	`head mystery.fasta` or `head ./mystery.fasta`

Optional challenge.

A. Vitis_vinifera.12Xv2.dna.toplevel.fa is 472 MB with 20 sequences (grep -c "^>"). head and grep -c both return in well under a second even on a file this size. Opening it in a text editor would try to load the whole 472 MB into memory at once, instead of reading a few bytes at a time.
B. For example, grep -c "^>" mystery.fasta on the first line and tail -n 2 mystery.fasta on the second, saved as answers.sh and run with bash answers.sh.

1.8 Before the next session

Keep the bioinfo_course folder. Session 2 builds a fuller project structure inside it and looks at FASTQ files and quality control.

Please finish your worksheet before Session 2

If you only got as far as copying mystery.fasta and WORKSHEET.md into place today, that is a completely normal amount to reach in the time we had. Please work through the rest at home before the next session: blocks D to G in the practical above are exactly the same whether you do them in class or on your own. It is still not graded; the point is that everyone starts Session 2 from roughly the same place.

Sources and reuse

This session adapts material from tbadet/teaching (Bioinformatic_tools) (RStudio Server access, file transfer, the essential-commands walkthrough, and the “Going further” commands) and from crolllab/IBIOL-compute-resources (cluster description and connection instructions). Screenshots are from the tbadet course materials; the cluster photo is from the IBIOL repository.

--- title: "Session 1: Initial diagnostic and essential commands" --- ## Overview [Download the lecture slides (PDF)](slides/01_lecture_slides.pdf) ::: callout-note ## Biological question What information can we recover from an unfamiliar sequence file? ::: This first session has two jobs. It gives everyone a shared mental model of **how a computer works** and **what a compute cluster is**, and it runs a short **diagnostic** so the rest of the course can be adjusted to the group. No prior experience is assumed. If you have used a terminal before, help the person next to you. ::: callout-tip ## Starting-point survey (about 4 min) Before the lecture, fill in the short survey: [starting-point survey](https://forms.gle/au3JoYFokMEEzbqx7), or scan the code below. ![](img/01/survey_qr.png){width="140"} It asks about your background, the tools you have already met, and the data you care about. Your name is **optional** and the survey is **not graded**: the answers are used to weight the examples for the rest of the course, and the class summary is shown back to you a few minutes later so you can see how mixed the starting levels are. This is separate from Section 1 of your worksheet, a self-assessment you fill in at the end of the practical (block G). Nothing in Sessions 1 and 2 is graded. ::: ### Learning outcomes By the end of this session, you should be able to: - name a few common types of biological data and say roughly how each is produced; - name the main parts of a computer and say what each one does; - explain what a compute cluster is and why we use one; - connect to the university cluster through RStudio Server (and know that `ssh` exists); - distinguish the RStudio **Console** from the **Terminal**; - identify your current folder and list its contents; - create and enter a folder; - copy a file and look inside it with `head`, `tail`, `grep`, and `wc`; - recognise **FASTA** records, count them reliably, and spot one that is not DNA or was truncated; - read an error message and correct the command; - keep a command log that another person can rerun from a clean start. ### Session plan | Time | Block | What happens | |---:|---|---| | 14:15-14:20 | Welcome | Anonymous starting-point survey (Form). | | 14:20-14:32 | Theory 1 | What bioinformatics is, and where biological data comes from. | | 14:32-14:45 | Theory 2 | How a computer works (live teardown). | | 14:45-15:00 | Theory 3-4 | From your laptop to a cluster. The FASTA format. Practical briefing. | | 15:00-15:15 | Practical A | Connect to RStudio Server. Console vs Terminal. Start a command log. | | 15:15-15:35 | Practical B | Navigate folders: predict each line, then run it. | | 15:35-15:50 | Practical C | Copy the file. Four ways to look inside it. | | 15:50-16:00 | Checkpoint | Save your work. Short break. | | 16:00-16:22 | Practical D | The investigation: what is wrong with `mystery.fasta`? (pairs) | | 16:22-16:40 | Practical E | Break it and diagnose: six failures, six causes. (pairs) | | 16:40-16:50 | Practical F | Reproduce a partner's script from an empty folder. | | 16:50-17:00 | Practical G | Individual wrap-up (not graded). | ## Theory (≈45 min) ### 1. Bioinformatics and the data it works on Bioinformatics connects a **biological question** to **data** and to a **reproducible analysis** that produces a result you can explain. ```{mermaid} flowchart LR Q["Biological question (e.g. which genes respond to stress in this organism?)"] --> D["Data (files: FASTA, FASTQ, tables…)"] D --> A["Analysis (a few commands, run in order)"] A --> R["Result (a number, a table, a figure)"] R --> I["Interpretation (answer + one limitation)"] ``` The analysis is just a **list of commands run in a defined order**. If you write the commands down, another person (including future you) can run them again and get the same result. That is what "reproducible" means, and it is the habit this course builds. #### Where does biological data come from? Different instruments answer different questions. You do not need to master any of these today: the goal is to recognise the main families and the kind of file each one lands in. | Data type | How it is produced | Typical files | Example question | In this course | |---|---|---|---|---| | Genome / amplicon sequencing | High-throughput sequencers (Illumina, Oxford Nanopore) read DNA fragments | FASTA, FASTQ | Which species are in this soil sample? | Sessions 2-3, 5-6 | | Bulk RNA-seq | Sequencing of RNA pooled from many cells → one expression value per gene | FASTQ → count table (CSV/TSV) | Which genes respond to drought? | Session 8 (project) | | Single-cell RNA-seq | RNA sequenced **per individual cell** | sparse matrix (MTX, H5) | Which cell types are present, and how do they differ? | mentioned only | | Proteomics | Mass spectrometry of peptides | mzML → abundance table | Which proteins change with treatment? | mentioned only | | Metabolomics | Mass spectrometry or NMR of small molecules | feature-intensity tables | Which metabolites mark a phenotype? | mentioned only | | Imaging / microscopy | Automated microscopes + image segmentation | TIFF / OME-TIFF → measurement tables | Does cell shape change between conditions? | mentioned only | | Genotyping / GWAS | SNP arrays or sequencing → genotype matrix | VCF, genotype tables | Which variants associate with a trait? | Session 6 / project | **The pattern to remember:** whatever the instrument, the analysis almost always ends up in one of two shapes of text file: **sequences** (FASTA, FASTQ) or **tables** (rows = samples or features, columns = variables). This course spends its time teaching you to open, check, and summarise those two shapes with confidence. **Bulk vs single-cell, in one line each.** *Bulk* measures the average over thousands of cells: cheap, and enough to ask "which genes differ between conditions". *Single-cell* resolves cell-to-cell variation and rare populations: more expensive, and the data are much sparser. How the sequencing itself works (library preparation, paired-end reads, read length) comes in Session 2, alongside FASTQ and quality control. ### 2. How a computer works A computer has a small number of parts. Each one has a single job. | Part | Job | Kitchen analogy | |---|---|---| | **CPU** (processor) | Does the calculations, one step after another (very fast). | The cook | | **RAM** (memory) | Holds the data currently being worked on. Fast, but **erased when the power goes off**. | The worktop: only what you are using right now fits on it | | **Disk / SSD** (storage) | Keeps files when the power is off. Slower than RAM, much bigger. | The fridge and cupboards | | **GPU** (graphics card) | Does many simple calculations in parallel. Used for images and for deep learning. | A brigade of line cooks all chopping at once | | **Operating system** (Linux, macOS, Windows) | Manages the parts, the files, and the programs. | The kitchen manager | | **Network card** | Talks to other computers. | The delivery door | Two ideas matter for the rest of the course: - **A program is just a file** that the CPU knows how to run. `ls`, `head`, `R` and `blast` are all programs sitting on the disk. - **Files live in folders, folders live in folders.** The chain of folder names that leads to a file is its **path**, for example `/home/you/bioinfo_course/data/reads.fasta`. The leading `/` is the top of the tree. The folder you are "standing in" right now is the **working directory**; a command that is given no path looks there. #### The shell The **shell** lets you drive the computer by typing commands instead of clicking. You type the name of a program, press <kbd>Enter</kbd>, and read what it prints back. If you know R, this is the same idea with a different language. The shell we use is called **bash**. It is good at three things bioinformatics needs constantly: navigating files, repeating an analysis exactly, and moving or summarising large files automatically. ### 3. From your laptop to a compute cluster A laptop is fine for writing code and looking at results. It is not fine for aligning a million sequencing reads: not enough CPU, not enough memory, and the data would not even fit. A **compute cluster** solves this. It is a set of powerful computers ("nodes") in a server room that many people share, with the large data stored right next to them. For this course we use **LEGcompute**, hosted by the Laboratory of Evolutionary Genetics and shared across the Institute of Biology ([crolllab/IBIOL-compute-resources](https://github.com/crolllab/IBIOL-compute-resources)). ![The LEGcompute cluster.](img/01/legcompute.jpeg){width="380"} | Resource | LEGcompute | |---|---| | CPU cores | ~1000 (Intel + AMD) | | GPUs | RTX 3090, 3× RTX A5500, 4× RTX 5000 Ada, 2× RTX PRO 6000 (96 GB) | | Memory | 4 TB RAM | | Storage | 80 TB SSD | | System | Ubuntu 26.04 LTS, Slurm scheduler, RStudio Server | #### How the cluster is organised ```{mermaid} flowchart LR L["Your laptop"] -- "https :8787 (RStudio)" --> H L -- "ssh" --> H H["Login / head node (where you land)"] -- "sbatch / srun" --> S["Slurm scheduler"] S --> C1["compute node"] S --> C2["compute node"] S --> G["GPU node"] H --- ST[("Shared storage /home /data /scratch")] C1 --- ST C2 --- ST G --- ST ``` - You **log in to the head node**. Everyone shares it, so it is for light interactive work only: editing, plotting, small commands (everything we do today). - Real jobs are **submitted to Slurm** with `sbatch`, which sends them to a free compute node so the machine is never overloaded and everyone gets a fair share. We meet Slurm properly in a later session. - **Storage is shared** by every node, so a file you create on the head node is visible in your jobs too: - `/home/username`: scripts and small files (**quota: 100 GB per student**); - `/data/username`: large datasets and analysis output (1+ TB); - `/scratch`: temporary space; **files older than 30 days are deleted automatically**. - **~200 bioinformatics tools** are pre-installed as *modules* (`module load SAMtools`). Nothing to install during class. ::: callout-important ## No backups No backup is made of anything on the server. Keeping a copy of work you care about is your responsibility. ::: #### Access for this course ::: callout-important ## Connecting to LEGcompute **1. Be on the university network.** Either the `unine` WiFi on campus, or the [GlobalProtect VPN](https://mydoc.unine.ch/fr/Internet_VPN_WiFi/VPN) from home. There is no way in from the open internet. **2. Claim your account.** [Open the shared Google Sheet the instructor](https://docs.google.com/spreadsheets/d/1QaN-ZJ-bWl8s_HbvMpZnnwgBn-5SQyBFRyZqk5vqST8/edit?usp=sharing). On the course tab, pick a free username and write your name in the yellow cell next to it. The first page of the sheet explains how to set or reset your password (via the [user account server](https://legserv.de6.quickconnect.to)). **3. Connect (pick one):** | Method | How | Good for | |---|---|---| | **RStudio Server** | Browser → <http://legcompute.unine.ch:8787> → log in. Use the **Terminal** tab for the shell. | **Everyone today.** Nothing to install. | | **SSH** | In a local terminal: `ssh username@legcompute.unine.ch` | Later, once you are comfortable. | We use **RStudio Server** for the whole course. It gives you an editor, an R Console, a Terminal and a file browser in one browser tab, with nothing to install. ::: ### 4. The FASTA format **FASTA** is the simplest way to store one or more biological sequences as text. A **record** has two parts: 1. a **header line**, whose first character is `>`, giving the sequence a name; 2. one or more **sequence lines** (the bases or amino acids). ```default >NC_045512.2 Severe acute respiratory syndrome coronavirus 2, surface glycoprotein region (partial) ATGTTTGTTTTTCTTGTTTTATTGCCACTAGTCTCTAGTCAGTGTGTTAATCTTACAACC AGAACTCAATTACCCCCTGCATACACTAATTCTTTCACACGTGGTGTTTATTACCCTGAC >oligo_F short PCR primer GATTACAGGCGTGAGCCACCACGT ``` Here the first record's sequence is wrapped over two lines; the second fits on one. Both are still single records. The next record starts at the next line beginning with `>`. Counting records is therefore the same as counting `>` lines. Nothing in the format tells you whether a sequence is DNA or protein, how long it is, or whether it is complete: you have to look. ### Practical briefing You will connect to the cluster, build a small project folder, and copy in one FASTA file, `mystery.fasta`. It is deliberately messy: different line widths, a record that is not DNA, and one that was cut off mid-download. Your job is to find that out with the commands, in pairs. **Write every command you run into a script file**, even (especially) the ones that fail: an error message is evidence, not a reason to start over. ## Practical (≈1 h 50) ### Essential commands for today Six commands to move around, plus four to look inside a file. That is the whole toolbox for today. | Command | What it does | Example | |---|---|---| | `pwd` | print working directory: "where am I?" | `pwd` | | `ls` | list the current folder (`ls -lh` adds sizes) | `ls -lh` | | `mkdir` | make a new folder | `mkdir data` | | `cd` | change directory: move into a folder | `cd data` | | `cp` | copy a file: `cp <what> <where>` | `cp reads.fasta .` | | `head` | show the first lines of a text file | `head -n 4 reads.fasta` | | `tail` | show the last lines | `tail -n 3 reads.fasta` | | `wc -l` | count lines | `wc -l reads.fasta` | | `grep "^>"` | show the lines that start with `>` (FASTA headers) | `grep "^>" reads.fasta` | | `grep -c "^>"` | *count* the lines that start with `>` (records) | `grep -c "^>" reads.fasta` | `^>` means "a line beginning with `>`". `cat` (whole file) and `less` (scroll, press `q` to quit) are useful too. Deleting files is deliberately **not** covered today. ::: callout-tip ## Three habits that save time - **Tab completion.** Type the first few letters of a command or file name and press <kbd>Tab</kbd>. The shell finishes it for you and prevents typos. - **Command history.** Press <kbd>↑</kbd> to bring back the previous command and edit it. - **Send a line from the script.** With your `.sh` script open, put the cursor on a line and press <kbd>Ctrl</kbd>+<kbd>Enter</kbd> (<kbd>Cmd</kbd>+<kbd>Enter</kbd> on macOS) to run it in the Terminal without copy-pasting. ::: ### A. Connect and get oriented (15:00-15:15) 1. In a browser, go to <http://legcompute.unine.ch:8787> and log in with your cluster username and password. ![The RStudio Server login page.](img/01/rstudio_login.png){width="360"} 2. Find these four parts of the RStudio window: - **Source editor** (top left): where scripts are written; - **Console** (bottom left): runs **R** code; - **Terminal** (a tab next to the Console): runs **shell** commands; - **Files** pane (bottom right): a click-through file browser. If there is no Terminal tab, open one with **Tools → Terminal → New Terminal**. The welcome text it prints is just a message from the server. ::: callout-warning ## Console vs Terminal Everything in this practical is typed in the **Terminal**. The **Console** is for R and comes in later sessions. If a shell command gives a strange R error, you are in the wrong panel. ::: 3. Open a new script: **File → New File → Shell Script**. Save it straight away as `session_01_commands.sh`. From now on, every command you try goes in this file, including ones that fail. Keep it: you reuse it in Session 2, and a partner reruns it in block F. ::: callout-note ## If you cannot connect Do not lose ten minutes alone. Pair up with a neighbour and share one screen, or use a room computer, and tell the instructor. Fixing access for one person should not stop the class. ::: ### B. Where am I? Navigating folders (15:15-15:35) Work **one line at a time**. Before you press <kbd>Enter</kbd>, write in your script, as a `#` comment, what you expect to see. Then run it and compare. A wrong prediction is the useful part: it shows exactly which idea to fix. ```bash pwd # predict: which folder does this print? ls # predict: what is already here? mkdir bioinfo_course # predict: what does ls show now? cd bioinfo_course pwd # predict: how has the path changed? mkdir session_01 cd session_01 pwd ``` ::: callout-note ## Checkpoint `pwd` ends in `bioinfo_course/session_01`. ::: Make a `data` folder and move into it. Predict each line first. ```bash mkdir data cd data pwd ``` ::: callout-note ## Checkpoint `pwd` ends in `bioinfo_course/session_01/data`. **Check this before block C**: the `cp` commands below copy into whatever folder you are currently standing in, so if `pwd` shows the wrong place, `cd` there first, or the files land somewhere you will not find them. ::: ### C. Get the data and take a first look (15:35-15:50) Make sure `pwd` still ends in `bioinfo_course/session_01/data` (you are copying files *into* this folder, not into the course materials folder), then copy the sequence file and your worksheet, and check they arrived: ```bash cp /home/cmaslard/bioinfo_tools_course/session_01/mystery.fasta . cp /home/cmaslard/bioinfo_tools_course/session_01/WORKSHEET.md . ls -lh ``` The final `.` means "here, in the current folder". `ls -lh` also shows the file sizes. `WORKSHEET.md` is a plain text file of questions, not a script: open it from the **Files** pane (click its name) or **File → Open File**, and type your answers directly under each "Answer:" line as you go through today's blocks. Save it like any other file (<kbd>Ctrl</kbd>/<kbd>Cmd</kbd>+<kbd>S</kbd>). There is nothing to submit: it stays in your folder, and your instructor may look at it to see where the class is. Now look inside it four ways. Predict the output of each line before you run it. ```bash head -n 6 mystery.fasta # the first few lines tail -n 3 mystery.fasta # the last few lines grep -c "^>" mystery.fasta # how many records (lines starting with >) wc -l mystery.fasta # how many lines in total ``` ::: callout-note ## Checkpoint `grep -c "^>"` prints **40**. `wc -l` prints **134**. Keep both numbers: block D asks why they differ. Note the output of `pwd` and the `ls -lh` line for `mystery.fasta` now, for **Section 2** of your worksheet. ::: ::: callout-tip ## Moving files in and out (no exercise, just so you know) You will not need it today, but the RStudio **Files** pane moves files between your laptop and the server: **Upload** (laptop to server) and tick a file then **More -> Export...** (server to laptop). No `scp` required. ![Uploading a file through the RStudio Files pane.](img/01/rstudio_upload.png){width="620"} ::: ### 15:50-16:00: Checkpoint and short break Before you step away: your script is **saved**, and `pwd` ends in `bioinfo_course/session_01/data`. ### D. The investigation, in pairs (16:00-16:22) `mystery.fasta` holds 40 records. Most are unremarkable filler; seven are not, and they are scattered in among the rest, not conveniently grouped at the top. Work with one other student. For **each** question below, write the answer **and the command you used** in Section 3 of your worksheet. 1. How many sequence records are in the file? Which command answers this without miscounting, and why is scrolling and counting `>` by eye unreliable here? 2. `wc -l` gave 134, not 80 (which would be two lines per record). Give the **two** separate reasons. 3. One record is **not DNA**. Which one? Give two clues, one from the sequence and one from the header. 4. One record was **cut off during download**. Which one, and how does `tail` show it? 5. Look at `contig_07`. Its sequence mixes UPPERCASE and lowercase and contains a run of `N`. What do the lowercase letters mean, and what does the `N` run mean? You do not need to change anything. 6. Which record has the **shortest complete sequence**? (The cut-off record from question 4 does not count: it has no sequence at all.) Show it and say how you decided. There is no single command that measures sequence length for you here: scroll through with `cat` or `less`, spot the records whose sequence fits on **one line**, and compare those by eye. ::: callout-note ## Short on time? At 16:22, move to block E with whatever you have, even if some questions are unanswered or your answers are unsure. Partial answers are fine: this is not graded, and there is no benefit in rushing block E to finish block D. ::: ::: {.callout-note collapse="true"} ## Commands you may want ```bash grep "^>" mystery.fasta # list every header at once grep -c "^>" mystery.fasta # count records head -n 20 mystery.fasta # a first chunk of records, to get a feel for the file tail -n 3 mystery.fasta # last lines grep -A 1 "SPIKE" mystery.fasta # a matching header and the line right after it ``` ::: ### E. Break it and diagnose, in pairs (16:22-16:40) Every command below fails or surprises you, each for a **different** reason. For each one: predict what will happen, run it, copy the **exact** message, then write a version that works. Record all of it in your script. ```bash cd Data head mystery.fna head -n mystery.fasta cp mystery.fasta cd mystery.fasta head ../mystery.fasta ``` Hints: - Linux distinguishes uppercase from lowercase. - A file name must match exactly, extension included. - Some options need a value; some commands need more than one argument. - `..` is "the folder above"; a bare name is "in the current folder". Return to `bioinfo_course/session_01/data` when you are done. ::: callout-note ## Short on time? At 16:40, move to block F even if you have not diagnosed all six. Whatever you have recorded, including the ones you did not solve, is still useful material for block G. ::: ### F. Reproduce it (16:40-16:50) Trade `session_01_commands.sh` with your partner. From your home folder, start clean and run **their** commands in order: ```bash cd ~ mkdir replay cd replay # now run your partner's commands, from the top, one by one ``` Does it run start to finish? If it stops, the missing step is something they did but did not write down, usually a `cd` or the `cp` path. Add the missing lines to **your own** script so it works from an empty folder. That is what "reproducible" means in practice. ### G. Individual wrap-up (16:50-17:00) On your own, complete Sections 1, 4, and 5 of your worksheet: - one command that failed today: the exact message, the cause, and the fix; - two or three sentences: from `mystery.fasta` alone, what **can** you say about these sequences, and what can you **not** say? You should leave with: - `WORKSHEET.md` with your answers; - `session_01_commands.sh` with every command you attempted, errors included, that runs from a clean start. Session 1 is a diagnostic, not a test. It is **not graded**: the continuous assessment (a short Moodle quiz at the end of each session) starts at Session 3. Your instructor may still look at your worksheet to see where the class is. **Keep the `bioinfo_course` folder.** Session 2 reuses it. ## Checkpoint You are done when all of the following are true: - you reached the Terminal in RStudio Server and know it is not the Console; - `session_01_commands.sh` is saved, includes the commands that failed, and runs from an empty folder; - `pwd` ends in `bioinfo_course/session_01/data`, and that folder contains `mystery.fasta`; - you can say the file has **40** records, that `grep -c "^>"` is the reliable way to count them, and why `wc -l` gives a different number; - you can name the record that is not DNA and the record that was truncated, and say how you spotted each; - you can explain in one sentence that a FASTA record starts on a line whose first character is `>`. ## Optional challenge For your own interest, not graded. **A. Look at a real genome without opening it.** LEGcompute keeps a shared folder of reference genomes that every account can read, no path from the instructor needed: ```bash ls -lh /data/genomes/fasta/ head /data/genomes/fasta/Vitis_vinifera.12Xv2.dna.toplevel.fa grep -c "^>" /data/genomes/fasta/Vitis_vinifera.12Xv2.dna.toplevel.fa ``` Note the file size, how many sequences `grep -c` finds, and how fast `head` answers even on a file of hundreds of megabytes. Why would opening it in the editor be a bad idea? **B. Make the answers reproducible.** Write a three-line script `answers.sh` that prints the record count and the last record of `mystery.fasta`, then run it with `bash answers.sh`. ## Going further: more shell commands ::: callout-tip ## Optional, if you have time left Not part of the essential route, not graded, and nothing later in the course assumes you have done it. If you reach 16:50 and blocks A to G are done, work through as much of this as you like; stopping partway through is completely fine. ::: A few more small, useful commands. Try each one in `bioinfo_course/session_01/data`. ### More ways to `ls` ```bash ls -l # long format: permissions, owner, size, date ls -a # include hidden files (names starting with a dot) ls -h # human-readable sizes, e.g. 6.4K instead of 6460 ls -lh # options can be combined ls -R # list this folder and every folder inside it ``` Run each one and compare. Which combination would you reach for to check a file's exact size before copying it somewhere with a tight quota? ### Writing and reading small text files `echo` prints text. On its own it writes to the screen; with `>` it writes to a file instead, replacing whatever was there. With `>>` it adds a line without erasing the rest. ```bash echo "first line" > notes.txt cat notes.txt # cat shows the whole file at once echo "second line" >> notes.txt cat notes.txt head -n 1 notes.txt # compare with cat: only the first line ``` `>` and `>>` are two characters apart and easy to mix up: `>` replacing a file you meant to add to is a classic way to lose work, so pause before you press Enter on one. ### Moving and renaming with `mv` `cp` leaves the original in place; `mv` does not; it moves or renames. ```bash mkdir backup cp mystery.fasta backup/ ls backup/ mv backup/mystery.fasta backup/mystery_copy.fasta # mv also renames ls backup/ ``` ### Two more worth knowing exist - `less mystery.fasta` scrolls a file a screen at a time with the arrow keys; press <kbd>q</kbd> to quit. Handy for a file too long for `head` to feel like enough. - `nano notes.txt` opens a small in-terminal text editor: type, <kbd>Ctrl</kbd>+<kbd>X</kbd> to exit, <kbd>y</kbd> to save. You already have an editor in RStudio's Source pane, so you may never need it here, but you will meet `nano` again on other servers. ## Answers ::: {.callout-note collapse="true"} ## Open this after you have tried the exercises yourself Session 1 is not graded, so there is no reason to hide the answers once you have genuinely tried each block. Use this to check your own reasoning, not instead of it: the value is in working out *why*, not in reading the line below. **Block C checkpoint.** `grep -c "^>" mystery.fasta` prints **40**; `wc -l mystery.fasta` prints **134**. **Block D, the investigation.** 1. **40 records**, from `grep -c "^>" mystery.fasta`. Counting `>` by eye is unreliable here because 40 headers are spread over 134 unevenly wrapped lines: easy to double-count or skip one while scrolling. 2. `wc -l` gives 134, not 80 (2 x 40), for two reasons: (a) most sequences wrap over two or more lines, adding lines beyond one per record; (b) `truncated_record` contributes only its header line, with no sequence line at all. 3. **Not DNA: `sp|P0DTC2|SPIKE_SARS2`.** Sequence clue: it contains letters that are not A/C/G/T/N (E, F, I, K, L, M, P, Q, R, S, V, W, Y). Header clue: `sp|...|` is a UniProt (Swiss-Prot) protein identifier, and the description says "glycoprotein". 4. **Truncated: `truncated_record`.** `tail -n 3 mystery.fasta` shows the file ending on a header line, with no sequence line after it. 5. In `contig_07`, lowercase letters mark **soft-masked** regions (repetitive sequence flagged by a masking tool, not removed). The run of `N` is an **assembly gap**: the length is estimated but the bases are unknown. 6. **Shortest complete sequence: `oligo_F`** (24 bases, one line). `truncated_record` has no sequence at all, so it counts as "cut off" (question 4), not "shortest". **Block E, six failures.** | Command | Fails because | A version that works | |---|---|---| | `cd Data` | case sensitivity: the folder is `data` | `cd data` | | `head mystery.fna` | wrong name: the file is `mystery.fasta` | `head mystery.fasta` | | `head -n mystery.fasta` | `-n` needs a number; it consumed `mystery.fasta` as the count | `head -n 5 mystery.fasta` | | `cp mystery.fasta` | `cp` needs a destination as well as a source | `cp mystery.fasta backup.fasta` | | `cd mystery.fasta` | `mystery.fasta` is a file, not a directory | `cd .` (stay), or `cd ..` to go up | | `head ../mystery.fasta` (from `data/`) | `..` is `session_01/`, where the file is not | `head mystery.fasta` or `head ./mystery.fasta` | **Optional challenge.** - **A.** `Vitis_vinifera.12Xv2.dna.toplevel.fa` is 472 MB with **20** sequences (`grep -c "^>"`). `head` and `grep -c` both return in well under a second even on a file this size. Opening it in a text editor would try to load the whole 472 MB into memory at once, instead of reading a few bytes at a time. - **B.** For example, `grep -c "^>" mystery.fasta` on the first line and `tail -n 2 mystery.fasta` on the second, saved as `answers.sh` and run with `bash answers.sh`. ::: ## Before the next session Keep the `bioinfo_course` folder. Session 2 builds a fuller project structure inside it and looks at FASTQ files and quality control. ::: callout-important ## Please finish your worksheet before Session 2 If you only got as far as copying `mystery.fasta` and `WORKSHEET.md` into place today, that is a completely normal amount to reach in the time we had. Please work through the rest at home before the next session: blocks D to G in the practical above are exactly the same whether you do them in class or on your own. It is still **not graded**; the point is that everyone starts Session 2 from roughly the same place. ::: --- ::: {.callout-note collapse="true"} ## Sources and reuse This session adapts material from [tbadet/teaching (Bioinformatic_tools)](https://github.com/tbadet/teaching/tree/main/Bioinformatic_tools) (RStudio Server access, file transfer, the essential-commands walkthrough, and the "Going further" commands) and from [crolllab/IBIOL-compute-resources](https://github.com/crolllab/IBIOL-compute-resources) (cluster description and connection instructions). Screenshots are from the tbadet course materials; the cluster photo is from the IBIOL repository. :::

1.1 Overview

1.1.1 Learning outcomes

1.1.2 Session plan

1.2 Theory (≈45 min)

1.2.1 1. Bioinformatics and the data it works on

1.2.1.1 Where does biological data come from?

1.2.2 2. How a computer works

1.2.2.1 The shell

1.2.3 3. From your laptop to a compute cluster

1.2.3.1 How the cluster is organised

1.2.3.2 Access for this course

1.2.4 4. The FASTA format

1.2.5 Practical briefing

1.3 Practical (≈1 h 50)

1.3.1 Essential commands for today

1.3.2 A. Connect and get oriented (15:00-15:15)

1.3.3 B. Where am I? Navigating folders (15:15-15:35)

1.3.4 C. Get the data and take a first look (15:35-15:50)

1.3.5 15:50-16:00: Checkpoint and short break

1.3.6 D. The investigation, in pairs (16:00-16:22)

1.3.7 E. Break it and diagnose, in pairs (16:22-16:40)

1.3.8 F. Reproduce it (16:40-16:50)

1.3.9 G. Individual wrap-up (16:50-17:00)

1.4 Checkpoint

1.5 Optional challenge

1.6 Going further: more shell commands

1.6.1 More ways to ls

1.6.2 Writing and reading small text files

1.6.3 Moving and renaming with mv

1.6.4 Two more worth knowing exist

1.7 Answers

1.8 Before the next session

1.6.1 More ways to `ls`

1.6.3 Moving and renaming with `mv`