Machine learning bridges omics sciences and plant breeding

Some of the biological knowledge obtained from fundamental research will be implemented in applied plant breeding. To bridge basic research and breeding practice, machine learning (ML) holds great promise to translate biological knowledge and omics data into precision-designed plant breeding. Here, we review ML for multi-omics analysis in plants, including data dimensionality reduction, inference of gene-regulation networks, and gene discovery and prioritization. These applications will facilitate understanding trait regulation mechanisms and identifying target genes potentially applicable to knowledge-driven molecular design breeding. We also highlight applications of deep learning in plant phenomics and ML in genomic selection-assisted breeding, such as various ML algorithms that model the correlations among genotypes (genes), phenotypes (traits), and environments, to ultimately achieve data-driven genomic design breeding.

The peripheral whole-blood transcriptome of acute pyelonephritis in human pregnancya

OBJECTIVE

Human pregnancy is characterized by activation of the innate immune response and suppression of adaptive immunity. The former is thought to provide protection against infection for the mother, and the latter, tolerance against paternal antigens expressed in fetal cells. Acute pyelonephritis is associated with an increased risk of acute respiratory distress syndrome and sepsis in pregnant (vs. nonpregnant) women. The objective of this study was to describe the gene expression profile (transcriptome) of maternal whole blood in acute pyelonephritis.

METHOD

A case-control study was conducted to include pregnant women with acute pyelonephritis (n=15) and women with a normal pregnancy (n=34). Affymetrix HG-U133 Plus 2.0 arrays (Affymetrix, Santa Clara, CA, USA) were used for gene expression profiling. A linear model was used to test the association between the presence of pyelonephritis and gene expression levels while controlling for white blood cell count and gestational age. A fold change of 1.5 was considered significant at a false discovery rate of 0.1. A subset of differentially expressed genes (n=56) was tested with real-time quantitative reverse transcription-polymerase chain reaction (qRT-PCR) (cases, n=19; controls, n=59). Gene ontology and pathway analyses were applied.

RESULTS

A total of 983 genes were differentially expressed in acute pyelonephritis: 457 were upregulated and 526 were downregulated. Significant enrichment of 300 biological processes and 63 molecular functions was found in pyelonephritis. Significantly impacted pathways in pyelonephritis included (a) cytokine-cytokine receptor interaction, (b) T-cell receptor signaling, (c) Jak-STAT signaling, and (d) complement and coagulation cascades. Of 56 genes tested by qRT-PCR, 48 (85.7%) had confirmation of differential expression.

CONCLUSION

This is the first study of the transcriptomic signature of whole blood in pregnant women with acute pyelonephritis. Acute infection during pregnancy is associated with the increased expression of genes involved in innate immunity and the decreased expression of genes involved in lymphocyte function.

Microarray profiling reveals that placental transcriptomes of early-onset HELLP syndrome and preeclampsia are similar

BACKGROUND

The involvement of the placenta in the pathogenesis of preeclampsia and HELLP syndrome is well established, and placental lesions are also similar in these two syndromes. Here we aimed to examine the placental transcriptome and to identify candidate biomarkers in early-onset preeclampsia and HELLP syndrome.

METHODS

Placental specimens were obtained at C-sections from women with early-onset preeclampsia and HELLP syndrome, and from controls who delivered preterm or at term. After histopathological examination, fresh-frozen placental specimens were used for microarray profiling and validation by qRT-PCR. Differential expression was analysed using log-linear models while adjusting for gestational age. Gene ontology and pathway analyses were used to interpret gene expression changes. Tissue microarrays were constructed from paraffin-embedded placental specimens and immunostained.

RESULTS

Placental gene expression was gestational age-dependent among preterm and term controls. Out of the 350 differentially expressed genes in preeclampsia and 554 genes in HELLP syndrome, 224 genes (including LEP, CGB, LHB, INHA, SIGLEC6, PAPPA2, TREM1, and FLT1) changed in the same direction (elevated or reduced) in both syndromes. Many of these encode proteins that have been implicated as biomarkers for preeclampsia. Enrichment analyses revealed similar biological processes, cellular compartments and biological pathways enriched in early-onset preeclampsia and HELLP syndrome; however, some processes and pathways (e.g., cytokine-cytokine receptor interaction) were over-represented only in HELLP syndrome.

CONCLUSION

High-throughput transcriptional and tissue microarray expression profiling revealed that placental transcriptomes of early-onset preeclampsia and HELLP syndrome largely overlap, underlying a potential common cause and pathophysiologic processes in these syndromes. However, gene expression changes may also suggest a more severe placental pathology and pronounced inflammatory response in HELLP syndrome than in preeclampsia.