Discovering transcription and splicing networks in myelodysplastic syndromes.

More and more transcription factors and their motifs have been reported and linked to specific gene expression levels. However, focusing only on transcription is not sufficient for mechanism research. Most genes, especially in eukaryotes, are alternatively spliced to different isoforms. Some of these isoforms increase the biodiversity of proteins. From this viewpoint, transcription and splicing are two of important mechanisms to modulate expression levels of isoforms. To integrate these two kinds of regulation, we built a linear regression model to select a subset of transcription factors and splicing factors for each co-expressed isoforms using least-angle regression approach. Then, we applied this method to investigate the mechanism of myelodysplastic syndromes (MDS), a precursor lesion of acute myeloid leukemia. Results suggested that expression levels of most isoforms were regulated by a set of selected regulatory factors. Some of the detected factors, such as EGR1 and family, are highly correlated with progression of MDS. We discovered that the splicing factor SRSF11 experienced alternative splicing switch, and in turn induced different amino acid sequences between MDS and controls. This splicing switch causes two different splicing mechanisms. experiments also confirmed that one of its isoforms was over-expressed in MDS. We analyzed the regulatory networks constructed from the co-expressed isoforms and their regulatory factors in MDS. Many of these networks were enriched in the herpes simplex infection pathway which involves many splicing factors, and pathways in cancers and acute or chronic myeloid leukemia.

KEYWORDS: {{ getKeywords(articleDetailText.words) }}