Gene Ontology: Pitfalls, Biases, and Remedies

Gaudet, Pascale; Dessimoz, Christophe

doi:10.1007/978-1-4939-3743-1_14

Cited by 144 publications

(109 citation statements)

References 34 publications

Supporting

Mentioning

106

Contrasting

Unclassified

Order By: Relevance

“…An additional "sliding window analysis," where a window of a certain length slides along the genotypes, checks whether SNPs under selection cluster in certain genomic regions (Tajima, 1991). After obtaining the regions under selection and the genes located there, we conduct a gene ontology (GO) enrichment analysis to explore which functional groups (GO terms) are over-represented for a specific gene set (Gaudet & Dessimoz, 2017;Primmer, Papakostas, Leder, Davis, & Ragan, 2013). In GO databases, the genes are assigned to predefined functional groups.…”

Section: Introductionmentioning

confidence: 99%

Genetic and phenotypic responses to genomic selection for timing of breeding in a wild songbird

et al. 2019

View full text Add to dashboard Cite

The physiological mechanisms underlying avian seasonal timing of reproduction, a life‐history trait with major fitness consequences, are not well understood. Comparing individuals that have been selected to differ in their timing of breeding may prove to be a promising in studying these mechanisms, making selection lines a valuable tool. We created selection lines for early and late timing of breeding in great tits (Parus major) using genomic selection, that is selection based on multi‐marker genotypes rather than on the phenotype. We took in nestlings (F1 generation) from wild broods of which the mother was either an extremely early (“early line”) or extremely late (“late line”) breeder. These chicks were then genotyped and, based on their “genomic breeding values” (GEBVs), we selected individuals for early and late line breeding pairs to produce the F2 generation in captivity. The F2 offspring was hand‐reared, genotyped and selected to produce an F3 generation, which were then again genotyped and selected. This way we obtained laying dates in aviaries for F1, F2 and F3 birds. We studied the genetic response to the artificial selection and found increased genetic differentiation between the early and late reproducing selection lines over generations (F1–F3), indicated by both diverging GEBVs and increased fixation indices (FST). We studied the phenotypic response to selection for birds breeding in outdoor breeding aviaries. We found that early line birds laid earlier than late line birds, and this difference increased over the generations (F1–F3), with non‐significant line effects for the F1 and F2, but highly significant line differences for the F3. We also assessed whether there was correlated selection on two traits that are potentially part of the mechanisms underlying seasonal timing: the endogenous free‐running period of the day/night clock (tau) and basal metabolic rate, but found no correlated selection. We have successfully created selection lines on seasonal timing in a wild bird species and obtained an instrument for future studies to investigate the physiological mechanisms underlying timing of breeding, and the genetic variation in these mechanisms, an essential component for evolutionary change in timing of reproduction. A free Plain Language Summary can be found within the Supporting Information of this article.

show abstract

Section: Introductionmentioning

confidence: 99%

Genetic and phenotypic responses to genomic selection for timing of breeding in a wild songbird

et al. 2019

View full text Add to dashboard Cite

show abstract

“…However, despite the efforts, at times the association of ontology and terms with genes is incomplete. This inherent incompleteness hinders the evaluation of computational methods and should be acknowledged when using the database . In our case, this means that some elements under GO:0003700 could include genes not specifically classified as TFs.…”

Section: Discussionmentioning

confidence: 99%

Integrated transcriptomics reveals master regulators of lung adenocarcinoma and novel repositioning of drug candidates

Bastiani

Klamt

2019

Cancer Medicine

View full text Add to dashboard Cite

Background Lung adenocarcinoma is the major cause of cancer‐related deaths in the world. Given this, the importance of research on its pathophysiology and therapy remains a key health issue. To assist in this endeavor, recent oncology studies are adopting Systems Biology approaches and bioinformatics to analyze and understand omics data, bringing new insights about this disease and its treatment. Methods We used reverse engineering of transcriptomic data to reconstruct nontumorous lung reference networks, focusing on transcription factors (TFs) and their inferred target genes, referred as regulatory units or regulons. Afterwards, we used 13 case‐control studies to identify TFs acting as master regulators of the disease and their regulatory units. Furthermore, the inferred activation patterns of regulons were used to evaluate patient survival and search drug candidates for repositioning. Results The regulatory units under the influence of ATOH8, DACH1, EPAS1, ETV5, FOXA2, FOXM1, HOXA4, SMAD6, and UHRF1 transcription factors were consistently associated with the pathological phenotype, suggesting that they may be master regulators of lung adenocarcinoma. We also observed that the inferred activity of FOXA2, FOXM1, and UHRF1 was significantly associated with risk of death in patients. Finally, we obtained deptropine, promazine, valproic acid, azacyclonol, methotrexate, and ChemBridge ID compound 5109870 as potential candidates to revert the molecular profile leading to decreased survival. Conclusion Using an integrated transcriptomics approach, we identified master regulator candidates involved with the development and prognostic of lung adenocarcinoma, as well as potential drugs for repurposing.

show abstract

“…However, our enrichment stats might still be biased by the chosen approach. Nevertheless, since GO annotations are dynamic and always biased by database representation 62 , we have chosen to keep the same methodological approach for both species. In this manner, if the enrichment test is biased it will be equally biased in both species facilitating comparisons.…”

Section: Discussionmentioning

confidence: 99%

Multiple lineages, same molecular basis: task specialization is commonly regulated across all eusocial bee groups

Araujo

Arias

2020

Preprint

View full text Add to dashboard Cite

14A striking feature of advanced insect societies is the existence of workers that forgo 15 reproduction. Two broad types of workers exist in eusocial bees: nurses which care 16 for their young siblings and the queen, and foragers who guard the nest and forage for 17 food. Comparisons between this two worker subcastes have been performed in 18 honeybees, but data from other bees are scarce. To understand whether similar 19 molecular mechanisms are involved in nurse-forager differences across distinct 20 species, we compared gene expression and DNA methylation profiles between nurses 21 and foragers of the buff-tailed bumblebee Bombus terrestris and of the stingless bee 22Tetragonisca angustula. These datasets were then discussed comparatively to 23 previous findings on honeybees. Our analyses revealed that although the expression 24 pattern of genes is often species-specific, many of the biological processes and 25 molecular pathways involved are common. Moreover, DNA methylation and gene 26 expression correlation were dependent on the nucleotide context. 27 subcastes evolved it is necessary to differentiate such more recent changes -that 53 could be species-specific -from those shared across species, and thus likely ancestral. 54The highly eusocial stingless bees have age-based division of labour 18 , similarly to 55 that of honeybees despite their most common ancestor being 50 to 80 million years 56 ago 19,20 . To date, no global expression or epigenetic studies have been performed in 57 stingless bees to understand worker task specialization. Similarly, while primitively 58 eusocial bumblebees are largely studied ecological biological models and important 59 wild and managed pollinators, we know comparatively little about the molecular 60 underpinnings of differences between its worker subcastes. Indeed, studies have been 61 restricted to few genes, leaving many open questions [21][22][23] . A major limiting element 62 for these studies is that this species display a somewhat fluctuating division of labour 63 with indistinctive separation between subcastes 10,21,23 . 64We aim to fill in this knowledge gap through the analyses of the global gene 65 expression differences between nurses and foragers, and the characterization of nurses 66 DNA methylation profile in two eusocial bee species, the primitively eusocial buff-67 tailed bumblebee, Bombus terrestris, and the highly eusocial stingless bee, 68Tetragonisca angustula. Combined, these two bee species and the honeybee represent 69 the three evolutionary branches of eusocial corbiculates sharing a common social 70 origin 24 . Hence, in addition to using the generated datasets to uncover unique and 71 more recent molecular traits linked to task division in B. terrestris and T. angustula, 72 we also verified whether common genes and pathways could be involved in task 73 specialization across all the eusocial bee groups. 74 75Results 76 task-dependent foraging gene expression in the bumblebee Bombus terrestris. Arch. Insect

show abstract

Gene Ontology: Pitfalls, Biases, and Remedies

Cited by 144 publications

References 34 publications

Genetic and phenotypic responses to genomic selection for timing of breeding in a wild songbird

Genetic and phenotypic responses to genomic selection for timing of breeding in a wild songbird

Integrated transcriptomics reveals master regulators of lung adenocarcinoma and novel repositioning of drug candidates

Multiple lineages, same molecular basis: task specialization is commonly regulated across all eusocial bee groups

Contact Info

Product

Resources

About