Synthetic-RNA-Seq-Raw-Counts-Generator

Generates synthetic Bulk RNA-Seq raw count data in a easy to use Google Colab Sheet, accelerated using vectorization & CUDA Acceleration

Synthetic Bulk RNA-Seq Dataset Generation Script to generate synthetic bulk RNA-seq count data with batch effects for batch correction evaluation. An optimized version that vectorizes the count generation to improve performance.

This workflow is near instantaneous at generating data.

Parameters:

• n_genes (int): Number of genes to simulate.
• n_samples (int): Total number of samples (should be divisible by n_batches).
• n_batches (int): Number of batches.
• de_fraction (float): Fraction of genes that are differentially expressed in treatment samples.
• fold_change (float): Fold change applied to the mean of differentially expressed genes in treatment.
• dispersion (float): Dispersion parameter for the negative binomial distribution.
• batch_effect_sigma (float): Standard deviation for batch effects (log-normal sigma).
• seed (int): Random seed for reproducibility.

Output:

pd.DataFrame: Count matrix with gene IDs as rows and sample IDs as columns. pd.DataFrame: Sample metadata with condition and batch information.

Contact

For more information https://github.com/stef1949 :3