Grouping data using Galaxy

Galaxy has numerous tools to analyse tables and variables they contains. Here, we are going use the tool Group data by a column and perform aggregate operation on other columns. (Galaxy Version 2.1.4) to rapidly extract information from a complex GTF file.

The input data

The input data is a GTF file annotation of the dmel 6.59 D. melanogaster genome which contains only annotations for exons on chromosomes X, 2, 3 and 4. This file is available on the GitHub repository ARTbio/AnalyseGenome at this URL:

https://github.com/ARTbio/AnalyseGenome/raw/refs/heads/main/Exercises/dmel.gtf.gz

1. Create a new history Grouping exercise, open the Upload menu (top-left corner, just below the Galaxy logo).

   Click the Paste/Fetch data button, and paste the above URL in the central field of the panel. Instead of the pre-filled New File, type dmel.gtf, and click the Start button.

   • Once the upload is complete, you will notice that the dataset's data type is gtf, not gtf.gz. This is because Galaxy uncompresses most uploaded data on the fly. Notable exceptions to this default behavior are fastq.gz and fasta.gz files. Since these files are often very large, they are kept in their compressed form to save disk space on the server and can be processed directly.

   • Have a look to the content of the file. Standard GTF ! 9 columns, the last column contains 4 types of annotations: gene_id, gene_symbol, transcript_id and transcript_symbol.

2. The first thing we are going to do is simplifying the 9^th column, keeping only the gene_id information.

   • Select the Column Regex Find And Replace in the left tools bar.
   • Select the column 9 for the dataset 1: dmel.gtf
   • Click the Insert Check
   • In the Find Regex enter the regular expression

     ;.+
     

   • In the Replacement field, do not enter anything
   • Click the RunButton.
   • In the resulting dataset 2: Column Regex Find And Replace on data 1, check the column 9 (Attributes). What do you see ?

3. First Grouping operation.

   Now, we are going to group the data in the 9^th column (we collapse lines by unique gene_id), while counting the "events" on column 3 (exon), and randomly picking a value on column 1 (the chromosome name, which is expected to be the same for all exons of a given gene)

   • Select the Group data by a column and perform aggregate operation on other columns in the left tools bar.
   • The Select data should already be 2: Column Regex Find And Replace on data 1
   • In the Group by column menu, select the 9^th column.
   • Click once on the Insert Operation
   • In the Type menu, select Count
   • In the On column menu, select Column: 3
   • Click one more time the Insert Operation
   • In the Type menu, select Randomly pick (bottom of the menu)
   • In the On column menu, select Column: 1
   • Click the RunButton !
   • What do you see in the generated dataset ? → It is a 3-column table. The first column contains unique gene_ids. The second column contains the computed number of exons for this gene. The third column contains the chromosome name for this gene.

4. Second Grouping operation.

   Finally, we are going the group again the data. This time, we group on the third column (chromosome names) while counting the variable (gene_id) in the 1 column.

   • Select the Group data by a column and perform aggregate operation on other columns in the left tools bar.
   • The Select data should already be 3: Group on data 2
   • In the Group by column menu, select the 3^rd column.
   • Click once on the Insert Operation
   • In the Type menu, select Count
   • In the On column menu, select Column: 1
   • Click the RunButton !

   • What do you see in the generated dataset ? → It is a 2-column table. The first column contains unique chromosome arms. The second column contains the computed number of genes for the chromosome arm.

     ?