【생물정보학】 NicheNet 이해하기

NicheNet 이해하기

 

추천글 : 【생물정보학】 리간드-수용체 상호작용 분석 

---

1. 개요 [본문]

2. 데이터셋 [본문]

3. 알고리즘 [본문]

4. 결론 [본문]

---

 

1. 개요 [목차]

⑴ 경로 활성화를 식별하고 관련된 잠재적인 리간드를 강조하는 툴

⑵ 즉, 특정 DEG가 있을 때 어떤 cell 혹은 어떤 ligand가 관련 있는지 구하는 것

⑶ 모식도 : 여러 데이터셋을 그래프 구조로 이어서 분석

 

출처 : 이미지 클릭

Figure. 1. NicheNet 모식도

 

⑷ 코드 : https://github.com/saeyslab/nichenetr

 

 

2. 데이터셋 [목차]

⑴ 종합 

① cpdb 

○ cpdb_interaction : Kamburov et al. (2013)

○ cpdb_complex : Kamburov et al. (2013)

② evex

○ lr_evex_regulation_expression : Van Landeghem et al. (2012)

○ evex_regulation_expression : Van Landeghem et al. (2012)

○ evex_catalysis : Van Landeghem et al. (2012)

○ evex_regulation_other : Van Landeghem et al. (2012)

○ evex_phosphorylation : Van Landeghem et al. (2012)

○ evex_regulation_binding : Van Landeghem et al. (2012)

○ evex_binding : Van Landeghem et al. (2012)

③ guide2pharmacology

○ pharmacology : Pawson et al. (2014)

④ harmonizome

○ harmonizome_KEA : Lachmann and Ma’ayan (2009)

○ harmonizome_PhosphoSite : Hornbeck et al. (2015)

○ harmonizome_kinase_substrate_predictions : Rouillard et al. (2016)

○ harmonizome_DEPOD : Duan et al. (2015)

○ harmonizome_CHEA : Lachmann et al. (2010)

○ harmonizome_ENCODE : Consortium (2004)

○ harmonizome_JASPAR : Mathelier et al. (2014)

○ harmonizome_TRANSFAC_CUR : Matys et al. (2006)

○ harmonizome_TRANSFAC : Matys et al. (2006)

○ harmonizome_MOTIFMAP : Xie et al. (2009)

○ harmonizome_GEO_TF : Edgar et al. (2002)

○ harmonizome_GEO_KINASE : Edgar et al. (2002)

○ harmonizome_GEO_GENE : Edgar et al. (2002)

○ harmonizome_MSIGDB_GENE : Subramanian et al. (2005)

⑤ HTRIDB 

○ HTRIDB : Bovolenta et al. (2012)

⑥ inweb_inbiomap

○ inweb_inbio_interaction : Li et al. (2017)

○ inweb_inbio_interaction_pathway : Li et al. (2017)

○ inweb_inbio_pathway : Li et al. (2017)

⑦ kegg

○ kegg_cytokines : Kanehisa et al. (2016)

○ kegg_cams : Kanehisa et al. (2016)

○ kegg_neuroactive : Kanehisa et al. (2016)

○ kegg_ecm : Kanehisa et al. (2016)

⑧ omnipath

○ omnipath_directional : Türei et al. (2016)

○ omnipath_undirectional : Türei et al. (2016)

⑨ ontogenet

○ ontogenet_coarse : Jojic et al. (2013)

⑩ pathwaycommons

○ lr_pathwaycommons_controls_expression_of : Cerami et al. (2011)

○ pathwaycommons_controls_expression_of : Cerami et al. (2011)

○ pathwaycommons_controls_phosphorylation_of : Cerami et al. (2011)

○ pathwaycommons_controls_state_change_of : Cerami et al. (2011)

○ pathwaycommons_catalysis_precedes : Cerami et al. (2011)

○ pathwaycommons_controls_transport_of : Cerami et al. (2011)

○ pathwaycommons_interacts_with : Cerami et al. (2011)

○ pathwaycommons_in_complex_with : Cerami et al. (2011)

⑪ ppi 

○ ppi_lr

○ ppi_l_bidir

○ ppi_bidir_r

○ ppi_bidir_bidir

○ ppi_lr_go

○ ppi_l_bidir_go

○ ppi_bidir_r_go

○ ppi_bidir_bidir_go

⑫ ramilowski

○ ramilowski_known : Ramilowski et al. (2015)

⑬ regnetwork

○ regnetwork_source : Liu et al. (2015)

○ regnetwork_encode : Liu et al. (2015)

⑭ Remap

○ Remap_5 : Griffon et al. (2015)

⑮ trrust

○ trrust : Han et al. (2015)

⑯ vinayagam

○ vinayagam_ppi : Vinayagam et al. (2011)

⑵ 종류 2. protein-protein interaction

① Omnipath

② PathwayCommons

③ InWeb

④ ConsensusPathDB

⑤ Vinayagam et al

⑥ EVEX

⑦ KEA

⑧ PhosphoSite

⑨ DEPOD

⑩ Harmonizome kinase–substrate predictions

⑶ 종류 3. gene regulatory interaction

① RegNetwork

② TRRUST

③ HTRIDB

④ ReMap

⑤ EVEX

⑥ PathwayCommons

⑦ Ontogenet

⑧ CHEA

⑨ ENCODE

⑩ JASPAR

⑪ TRANSFAC

⑫ MOTIFMAP

⑬ Gene Expression Omnibus (GEO)

⑭ MSigDBHarmonizome

 

 

3. 알고리즘 [목차]

⑴ 단계 1. 한 개의 통합 네트워크가 아니라 두 개의 네트워크로 진행

① 종류 1. protein–protein interaction for ligand–receptor and signaling data sources

② 종류 2. gene regulatory interaction in the gene regulatory network

⑵ 단계 2. weighted sum of adjacency matrices

 

 

⑶ 단계 3. mlrMBO를 이용한 parameter optimization

① mlrMBO : modular framework for model-based optimization of expensive black-box functions

⑷ 단계 4. weighted matrix를 랜덤하게 생성

⑸ 단계 5. ligand-target regulatory potential score 계산

 

 

⑹ 단계 6. 모델 생성

① 목적 : 다중 data source 중 하나를 선별하기 위함 

② 종류 1. leave-one-in

③ 종류 2. one-versus-one-versus-one

⑺ 단계 7. validation 

① target gene prediction 평가

② ligand activity prediction 평가

③ cell type bias 평가

④ IPA(upstream regulator analysis of ingenuity pathway analysis)와의 비교

⑤ CCCExplorer와의 비교

⑻ 단계 8. 다른 데이터셋에서의 적용 및 비주얼라이제이션

 

 

4. 결론 [목차]

⑴ CellPhonedB보다 더 좋다는 의견이 있음 

 

입력: 2023.09.15 11:15