Table 5.
Functional clusters among target genes or their regulatory candidates
| Regulated targets a | Regulatory candidates b |
|---|---|
|
Fatty acid metabolism
c
| |
| Cluster 1: ACAA2, ACADL, ACADVL, ACSS2, ALDH9A1, CLU, DECR1, DGAT2, DHRS4, ECH1, ARCC1, FNTB, HADHB, HMGCS2, LSS, MBLN3, NME4, PAPSS2, PDK1, PEX11A, PNPLA2, PYCR1, RETSAT, UCP2 |
Cluster 1: PNPLA2,PPARA, PPARG,SERPINA3, TGFB1 |
| Cluster 2: ACADL, ACADVL, ALDH9A1, DECR1, DHRS4, EGLN3, PYCR1, RETSAT |
Cluster 2: ACAA2, ACADL, ACADVL, ECH1, HADHB, PNPLA2,PPARA, PPARD |
|
Cholesterol metabolism | |
| Cluster 4: ACAA2, ACSS2, DGAT2, HMGCS2 |
Cluster 3: ACAA2, ACSS2, DGAT2, HMGCS2, LSS,PPARD, RXRA |
|
Cell adhesion and migration | |
| Cluster 7: CDH11, CLU, COL18A1, DPT, LAMA2, LAMA4, PTK7, SERPINE2 |
Cluster 4: COL18A1, LAMA2, LAMA4, SERPINE2, TGFB1 |
|
Cell development | |
| Cluster 3: SERPINA3, SERPINE1, SERPINE2 |
Cluster 6: CLU, COL18A1, DPT, EGLN3, FNTB,PPARD, PPARG, RXRA, SERPINE1, TGFB1 |
| |
Cluster 5: EGLN3, LAMA2, LAMA4,PPARD, PPARG, RXRA, RXRG, TGFB1 |
| |
Cluster 8: CLU, JUNB, JUND, PPARD, PPARG, TGFB1 |
|
Cell signaling | |
| Cluster 6: APCDD1, CA6, CDH11, CLU, COL18A1, DGAT2, DPT, HTRA3, LAMA2, LAMA4, LCN2, LGALS9, NPR3, PNPLA2, OPDC3, PTK7, RETSAT, SERPINA3, SERPINE1, SERPINE2, STRA6 |
Cluster 7: (PPAR signaling pathway) : ARNTL, ARNTL2, CLOCK, CLU, COL18A1, DPT, GLN3, ERCC1, FNTB, HLTF, HMGCS2, JUNB, JUND, LAMA2, LAMA4, MBNL3, PAPSS2, PNPLA2,PPARA, PPARD, PPARG, RXRA, RXRG,SERPINA3, SERPINE1, TGFB1 |
|
Peroxisome microbody | |
| Cluster 5: DHRS4, ECH1, PEX11A |
|
| Cluster 9: CDH11, JUND, NPR3, PAPSS2, TGFB1 | |
aTranscripts experimentally identified as responsive to PPARA agonists in cell culture [21] (see Additional file 2: Table S2 for details) were clustered using the DAVID functional annotation tool [5]. Only clusters with an enrichment > 0.5 and a Benjamini score < 0.15 were kept. Eventually, a synthetic description was chosen to name the cluster.
bFifty upstream regulatory candidates that were automatically-proposed according to best specificity scores (see Additional file 2: Table S2 for a detailed list) were clustered. All molecules were first linked to their related genes, with heterodimer protein complexes (e.g., PPARA:RXRA) being switched in the two genes (e.g., PPARA and RXRA).
cThe results show that the input list of regulated transcripts and the output list of proposed candidates notably shared clusters related to fatty acid and cholesterol metabolisms. All clusters were enriched in transcription factors, and especially in peroxisome proliferator-activated receptors (PPAR) isotypes and their partners RXR (indicated in bold face) when calculated from the list of regulatory candidates.