Choose the Right Files

Point Drylab at the exact files or folder before you ask, so it analyzes your data and not a guess.

When a project holds several versions of the same data, naming the exact input is the fastest way to get a correct answer. Mention the files or folder in your message, and Drylab uses exactly those.

When to use it

Your project has counts.csv, counts_filtered.csv and counts_v2.csv, and you want Drylab to use one of them.
You want to run one step on a whole sequencing run, such as every FASTQ file in a folder.
You want Drylab to read a figure or a table that's already in your project before it changes anything.
You're starting a new analysis and want Drylab to check the files before it runs anything.

Prompt examples

Check the inputs first:

@counts.csv and @metadata.csv Before any analysis, inspect both files: list the columns, the number of genes and samples, and check that the sample IDs match between the two files. Don't run any analysis yet.

Name the columns you mean:

@counts.csv and @metadata.csv Run a paired differential expression of tumor vs normal. Use the condition column for the groups and the patient column for the pairing.

A whole folder:

@o fastq_run_01 Run FastQC on every FASTQ file in this folder and give me one summary table with the read count and mean quality per sample.

A figure:

@figure1.png What does this plot show, and are there any problems with the axes, labels or colors?

One file out of several similar ones:

Use only @counts_filtered.csv, not the raw counts. Make a PCA of the samples colored by condition from @metadata.csv.

Steps

Choose the project first. On the home page, click the project name under the chatbox and pick your study. The @ picker only lists the files of the current project. See Projects.
Make sure your files are in the project. To add new ones, click + (Upload files or mention resources) in the chatbox and choose Upload file or Upload folder, or drag files or a folder onto the chatbox. They go to the project's Project Files folder. See Upload Files.
In the chatbox, type @ and start typing the file name. To list only files, type @f (with a space); for folders, type @o .
Pick the file with the arrow keys and Enter, or click it. It becomes a chip in your message. Repeat for each file you need. + > Use Resource / Tools opens the same picker.
Type the rest of your request, then send it.

Open the project folder that holds your data. In the Desktop App, a project is a folder on your computer. See Use the Desktop App.
To bring in a file from elsewhere on your computer, drag it onto the chatbox. When it shows "Drop files to attach", let go. Drylab adds the file to Project Files in your project folder without uploading it, and attaches it to your message.
To use a file that's already in the project, find it in the File tree in the left sidebar. Right-click it and choose Attach to Chat Box, or hover it and click the paperclip (Attach to chat).
Or type @ in the chatbox and start typing a name. In the Desktop App, @ searches your whole project folder, not only Project Files.
Type the rest of your request, then send it.

The chatbox + menu (Upload file, Upload folder) uploads to your cloud Vault instead. To keep your data on your computer, use drag and drop or the File tree. See Work with Local Files.

When Drylab asks you for a file

Sometimes Drylab stops and asks for data, for example "Upload the sample metadata file". Click Upload files to project on the question card, or drop files onto it, then click Submit Answers. See Questions that need a file.

Tips

Mention the data and the method together, in one message: "@counts.csv and @metadata.csv, paired DE", not two messages.
Name the columns too. Drylab is faster and more accurate when you say which column holds the groups, the pairing or the gene IDs.
Ask for a check first on new data. A quick "inspect the files, don't run any analysis yet" catches mismatched sample IDs before they cost you a run.
Keep one study per project, so the @ picker shows only the files that belong together.