Improving gene function predictions using independent transcriptional components

• Published in: Nature communications Vol. 12; no. 1; p. 1464
• Main Authors: Urzúa-Traslaviña, Carlos G., Leeuwenburgh, Vincent C., Bhattacharya, Arkajyoti, Loipfinger, Stefan, van Vugt, Marcel A. T. M., de Vries, Elisabeth G. E., Fehrmann, Rudolf S. N.
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 05.03.2021; Nature Publishing Group; Nature Portfolio
• Subjects: 631/114/2397; 631/114/2403; 631/114/2404; 631/114/2412; Animals; Computational Biology; Functional groups; Gene Expression Profiling - methods; Gene Knockout Techniques; High-Throughput Nucleotide Sequencing; Humanities and Social Sciences; Humans; Independent component analysis; Mice; multidisciplinary; Next-generation sequencing; Predictions; Principal components analysis; RNA, Messenger - metabolism; Science; Science (multidisciplinary); Sequences; Transcription; Transcriptome
• ISSN: 2041-1723; 2041-1723
• DOI: 10.1038/s41467-021-21671-w

The interpretation of high throughput sequencing data is limited by our incomplete functional understanding of coding and non-coding transcripts. Reliably predicting the function of such transcripts can overcome this limitation. Here we report the use of a consensus independent component analysis and guilt-by-association approach to predict over 23,000 functional groups comprised of over 55,000 coding and non-coding transcripts using publicly available transcriptomic profiles. We show that, compared to using Principal Component Analysis, Independent Component Analysis-derived transcriptional components enable more confident functionality predictions, improve predictions when new members are added to the gene sets, and are less affected by gene multi-functionality. Predictions generated using human or mouse transcriptomic data are made available for exploration in a publicly available web portal. Our understanding of the function of many transcripts is still incomplete, limiting the interpretability of transcriptomic data. Here the authors use consensus-independent component analysis, together with a guilt-by-association approach, to improve the prediction of gene function.