Skip to content

Load

Load external files into OmicsBox objects (Figure 1).

image-20240416-095145.png

Figure 1. Transcriptomics Load functions.

Load RNA-Seq Count Table

Load bulk RNA-Seq Count Tables from text files (Figure 2). The first line must contain the column (sample) names and the first column must contain gene/feature names (Figure 3).

  • Count Table File. Specify the count table file in .txt, .tsv, or .csv formats.
  • Column Separator. The character that separates the columns: tab (" "), space (" "), comma (","), or semicolon (";")

  • NA Values. How to handle missing values:

  • Skip Line: do not load the counts for the entire row (gene).

  • Assume Zero Values: replace the NA value with a 0.

image-20240416-094938.png

Figure 2. Load RNA-Seq Count Table wizard.

image-20240416-100238.png

Figure 3. Example RNA-Seq Count Table txt file, separated by tabs.

Load RNA-Seq Pairwise Results

Load a tab-delimited file containing the results from a pairwise differential expression analysis (Figure 4). The table must meet the following conditions:

  • It must contain the following mandatory fields: logFC, logCPM, PValue, and FDR.
  • It must be tab-delimited.
  • The type of information contained in each column must be indicated in its header (first line).
  • The first column must contain the name of each sequence (gene or transcript).
  • It may contain other column fields that are named according to their header.

image-20240416-102615.png

Figure 4. Example Pairwise Results file.

Load Single-cell RNA-Seq Count Matrix

Load Single-cell RNA-Seq (scRNA-Seq) count matrices in different formats into OmicsBox (Figure 5). An scRNA-Seq Count Matrix object will be opened in a new tab.

  • Input Type. Select the format of the count table:

  • Matrix Market File. The output of bioinformatics tools like CellRanger or STARsolo. It consists of three parts: an MTX file (.mtx) with the locations of non-zero counts, a barcodes file with cell IDs (column names), and a features file with gene names (row names). The MTX file must meet the specifications explained here. Some important points to note:

    • The barcodes and features files must not have a header.
    • The number of lines in the barcodes and features files must match the size specified in the MTX file (Figure 6).
    • The features file can have extra columns for feature metadata, separated by tabs. The first column should have Gene IDs, the second should have Gene Names, and any additional columns will be ignored.

    • Text File. Output of tools like RSEM, Drop-seq tools, etc. It is a text file similar to Figure 3, containing cells in columns and genes in rows.

    • Column Separator: Only enabled if Text File is selected. Specify the character that separates the columns: tab (" "), space (" "), comma (","), or semicolon (";")

    • CellRanger H5. The output of CellRanger with .h5 extension. Other count matrices in H5 are not supported. The CellRanger’s format is explained in detail here.
    • H5 Annotated Data. Another type of h5 files (with the extension .h5ad) that follow the AnnData format. This format stores the count matrix in a group named "/X", the cell metadata in a group named "/obs", and gene metadata in a group named "/var". Further specifications can be found here. It is the most commonly used format in databases containing annotated scRNA-Seq references.
    • Loom Matrix File. It is the output of tools like Kallisto+BUStools, zUMIs, etc. It is another type of h5 file with a specific format, detailed here. The matrix has to be stored in a group named "/matrix", the cell names in a group named "/col_attrs/CellID", and the gene names in a group named "/row_attrs/Gene".

Additional options will be incorporated as new formats are developed to address user requirements.

When loading counts in MTX, Text, and H5 Annotated Data, fractional counts will be turned into integers. These fractional counts are generated by algorithms like Salmon, which distribute the counts of multi-mapping reads across the mapped genes, leading to decimal counts.
However, many scRNA-Seq analysis tools only accept integer counts. So, these fractional counts are converted to integers by removing the decimal part, not by rounding. This is done to avoid inflating the data size with many 0.5 counts being rounded up to 1, which wouldn’t be a meaningful count.

image-20240416-103106.png

Figure 5. Load Single-cell RNA-Seq count matrix wizard.

image-20240416-110844.png

Figure 6. Example of a scRNA-Seq count matrix in MTX format.