The pathway not taken: understanding 'omics data in the perinatal context

Area of Science:

• Genomics and Bioinformatics
• Perinatal Research
• Developmental Biology

Background:

• Omics analysis is crucial in perinatal research, but interpreting fetal gene expression remains challenging.
• Existing systems biology resources like Ingenuity Pathway Analysis (IPA) may not fully capture fetal-specific biological context.
• Developmental Functional Annotation at Tufts (DFLAT) offers curated functional annotations for fetal development.

Purpose of the Study:

• To compare the interpretation of fetal transcriptome data using IPA versus GSEA with DFLAT.
• To assess the biological relevance of each analytical approach for fetal developmental perturbations.

Main Methods:

• Analyzed amniotic fluid transcriptome datasets from three fetal conditions: Trisomy 21 (T21), twin-twin transfusion syndrome (TTTS), and maternal obesity.
• Used paired t-tests with Benjamini-Hochberg correction to identify differentially expressed probe sets.
• Performed functional analyses using IPA and GSEA/DFLAT, comparing outputs for fetal relevance.

Main Results:

• Both IPA and GSEA/DFLAT identified common dysregulated pathways across T21, TTTS, and maternal obesity.
• GSEA/DFLAT provided more comprehensive functional insights, particularly for brain development and in cases of maternal obesity.
• IPA generated more annotations related to cell death and erroneously flagged physiological fetal proliferation as cancer-related.

Conclusions:

• The choice of analytical program impacts the interpretation of fetal transcriptome data, necessitating the use of multiple resources.
• IPA’s bias towards adult diseases can lead to false positives in fetal studies.
• Gene annotation resources with a developmental focus, such as DFLAT, are recommended for perinatal omics studies.