Explicit Scale Simulation for analysis of RNA-sequencing count data with ALDEx2

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• High-throughput sequencing (HTS) studies face technical variations affecting sequencing depth.
• Traditional normalization methods in HTS analysis make assumptions about biological system scale, potentially leading to errors.
• Errors in normalization assumptions can increase false positive and false negative rates in differential abundance analyses.

Purpose of the Study:

• To introduce and demonstrate the application of scale models in transcriptomic analysis.
• To show how scale models mitigate errors caused by normalization assumptions in HTS data.
• To enhance transparency and reproducibility in transcriptomic data analysis.

Main Methods:

• Integration of scale models into the ALDEx2 R package.
• Application of scale models to transcriptomic case studies, including metatranscriptomics.
• Utilizing known housekeeping genes to build scale models for complex datasets.

Main Results:

• Scale models reduce false positive and false negative rates compared to traditional normalizations in transcriptomic data.
• Scale models enhance the transparency and reproducibility of HTS data analyses.
• Scale models effectively address the disconnect between practical and statistical significance, replacing dual cutoff approaches.

Conclusions:

• Scale models offer a more robust approach to analyzing HTS data, particularly in transcriptomics.
• Incorporating scale into analysis mitigates technical variation and improves biological interpretation.
• This work provides practical guidance for applying scale models to transcriptomic datasets for more reliable results.