IA explicável aplicada para identificar genes influentes na classificação do câncer por meio de dados de expressão gênica de RNA-Seq DOI Open Access
Karolayne Azevedo,

Luísa Souza,

Matheus Dalmolin

et al.

Published: Dec. 31, 2023

Este artigo faz uso de três técnicas aprendizagem máquina (Machine Learnig – ML) para classificar os cinco tipos câncer mais recorrentes em mulheres, a partir dados expressão gênica RNA-Seq. Os desafios incluem: alta dimensionalidade do conjunto e falta transparência dos modelos ML. Para mitigar esses problemas, foi utilizado técnica SHAP (SHapley Additive exPlanations) que uma inteligência artificial explicável (Explainable intelligence XAI) utilizada compreender como tomam decisões podendo ser usada estratégia seleção recursos. Como entrada, foram utilizadas 2.105 amostras, sendo 421 amostras referentes cada tumor, processadas pelos Arvore Decisão (Decision Tree- DT), Floresta Aleatoria (Random Forest-RF) Aumento Gradiente Extremo (eXtreme Gradient Boosting-XGB) treinadas validadas por meio da validação cruzada. RF, DT XGB alcançaram precisões 99, 40%, 97, 60% 34%. Posteriormente, obter lista recursos visando quais características influenciaram nas tomadas consequentemente, nos resultados predição tumores. 122, 90 11 genes obtidos DT, totalizando 223 resultando 194 únicos.

ZBTB7A forms a heterodimer with ZBTB2 and inhibits ZBTB2 homodimerization required for full activation of HIF-1 DOI Creative Commons
Gouki Kambe, Minoru Kobayashi, Hiroshi Ishikita

et al.

Biochemical and Biophysical Research Communications, Journal Year: 2024, Volume and Issue: 733, P. 150604 - 150604

Published: Aug. 24, 2024

Hypoxia-inducible factor 1 (HIF-1), recognized as a master transcription for adaptation to hypoxia, is associated with malignant characteristics and therapy resistance in cancers. It has become clear that cofactors such ZBTB2 are critical the full activation of HIF-1; however, mechanisms downregulating ZBTB2-HIF-1 axis remain poorly understood. In this study, we identified ZBTB7A negative regulator by analyzing protein sequences structures. We found forms heterodimer ZBTB2, inhibits homodimerization necessary expression downstream genes, ultimately delays proliferation cancer cells under hypoxic conditions. The Cancer Genome Atlas (TCGA) analyses revealed overall survival better patients high their tumor tissues. These findings highlight potential targeting ZBTB7A-ZBTB2 interaction novel therapeutic strategy inhibit HIF-1 activity improve treatment outcomes hypoxia-related

Language: Английский

Citations

0
Thyroid gland: Anatomy and physiology DOI
Salvatore Benvenga, Giovanni Tuccari, Antonio Ieni

et al.

Elsevier eBooks, Journal Year: 2024, Volume and Issue: unknown

Published: Jan. 1, 2024

Language: Английский

Citations

0
Patterns of senescence and apoptosis during development of branchial arches, epibranchial placodes, and pharyngeal pouches DOI Creative Commons
Stefan Washausen, Wolfgang Knabe

Developmental Dynamics, Journal Year: 2023, Volume and Issue: 252(9), P. 1189 - 1223

Published: June 22, 2023

Abstract Background Many developmental processes are coregulated by apoptosis and senescence. However, there is a lack of data on the development branchial arches, epibranchial placodes, pharyngeal pouches, which harbor signaling centers. Results Using immunohistochemical, histochemical, 3D reconstruction methods, we show that in mice, senescence together may contribute to invagination clefts deepening cervical sinus floor, antagonism proliferation acting evaginating arches. The concomitant apoptotic elimination lateral line rudiments occurs absence In appear (1) support or at least indentation immobilizing margins centrally proliferating pit, (2) coregulate number fate Pax8 + precursors, (3) progressively narrow neuroblast delamination sites, (4) placode regression. Putative centers pouches likely deactivated rostral caudal apoptosis. Conclusions Our results reveal plethora novel patterns senescence, some overlapping, complementary, whose functional contributions region, including placodes their centers, can now be tested experimentally.

Language: Английский

Citations

1
Prognostic Value of Pax8 in Small Cell Lung Cancer DOI

Fengyun Tao,

Hangyan Zhu,

Jiayun Xu

et al.

Published: Jan. 1, 2023

Objectives: Small cell lung cancer (SCLC) shows poor prognosis since it metastasizes widely at early stage. PAX8 is a transcriptional factor that belongs to Paired box gene (PAX) family. Expression of in controversial issue. The prognostic value SCLC still unclear.Materials and Methods: Overall, 122 subjects who were pathologically diagnosed with enrolled the study. Immunohistochemical analysis Ki-67 performed. correlations between expression clinical features or index further analyzed. Subsequently, an association PAX8, stage, status, overall survival (OS) was performed 107 follow-up information.Results: positive 50% (61/122) specimens. rate significantly higher extensive-stage specimens (61.29%) than limited-stage (38.33%). level positively correlated (P=0.012) negatively OS (HR=5.255, 95% CI 1.724-16.012, P=0.004). In combination groups, negative limited stage group has most promising which no death during period.Conclusion: not low. It small cancer.

Language: Английский

Citations

0
IA explicável aplicada para identificar genes influentes na classificação do câncer por meio de dados de expressão gênica de RNA-Seq DOI Open Access
Karolayne Azevedo,

Luísa Souza,

Matheus Dalmolin

et al.

Published: Dec. 31, 2023

Este artigo faz uso de três técnicas aprendizagem máquina (Machine Learnig – ML) para classificar os cinco tipos câncer mais recorrentes em mulheres, a partir dados expressão gênica RNA-Seq. Os desafios incluem: alta dimensionalidade do conjunto e falta transparência dos modelos ML. Para mitigar esses problemas, foi utilizado técnica SHAP (SHapley Additive exPlanations) que uma inteligência artificial explicável (Explainable intelligence XAI) utilizada compreender como tomam decisões podendo ser usada estratégia seleção recursos. Como entrada, foram utilizadas 2.105 amostras, sendo 421 amostras referentes cada tumor, processadas pelos Arvore Decisão (Decision Tree- DT), Floresta Aleatoria (Random Forest-RF) Aumento Gradiente Extremo (eXtreme Gradient Boosting-XGB) treinadas validadas por meio da validação cruzada. RF, DT XGB alcançaram precisões 99, 40%, 97, 60% 34%. Posteriormente, obter lista recursos visando quais características influenciaram nas tomadas consequentemente, nos resultados predição tumores. 122, 90 11 genes obtidos DT, totalizando 223 resultando 194 únicos.

Citations

0