2026-09-18T20:16:12Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1567662024-09-17T03:02:43Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Kellis, Manolis
Yan, Binwei
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2024-09-16T13:47:50Z
2024-09-16T13:47:50Z
2024-05
2024-07-11T14:36:47.766Z
https://hdl.handle.net/1721.1/156766
This thesis investigates diverse computational methodologies for modeling cellular interactions using single-cell RNA sequencing (scRNA-seq) data. We evaluate the performance of Graph Neural Networks (GNNs) both with and without gene-gene edges, Contrastive Learning, and Variational Autoencoders (VAEs) across multiple datasets. Our study compares these methods and establishes benchmarks for assessing their effectiveness beyond traditional case studies. By integrating extensive signaling pathway data, we aim to unveil complex cell-cell communication patterns and regulatory mechanisms that conventional scRNA-seq analysis methods might overlook. Our approach emphasizes the use of spatial data as a crucial indicator, facilitated by the advanced capabilities of heterogeneous GNNs to model physical proximity. We found that our analysis of the functioning genes aligns with previous findings, proving our model’s effectiveness as a potential method for further analyze communication mechanisms.
M.Eng.
Massachusetts Institute of Technology
Attribution-NonCommercial-NoDerivatives 4.0 International (CC BY-NC-ND 4.0)
Copyright retained by author(s)
https://creativecommons.org/licenses/by-nc-nd/4.0/
Using heterogeneous Graph Neural Networks(hGNN) to predict cell-cell communication
Thesis
application/pdf
Master
Master of Engineering in Electrical Engineering and Computer Science
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="edb924c3-80bc-40ac-b07a-89c0586a7368">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Using heterogeneous Graph Neural Networks(hGNN) to predict cell-cell communication</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2024-05</PublicationDate>
<Authors>
<Author>
<DisplayName>Yan, Binwei</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://creativecommons.org/licenses/by-nc-nd/4.0/</License>
<Abstract>This thesis investigates diverse computational methodologies for modeling cellular interactions using single-cell RNA sequencing (scRNA-seq) data. We evaluate the performance of Graph Neural Networks (GNNs) both with and without gene-gene edges, Contrastive Learning, and Variational Autoencoders (VAEs) across multiple datasets. Our study compares these methods and establishes benchmarks for assessing their effectiveness beyond traditional case studies. By integrating extensive signaling pathway data, we aim to unveil complex cell-cell communication patterns and regulatory mechanisms that conventional scRNA-seq analysis methods might overlook. Our approach emphasizes the use of spatial data as a crucial indicator, facilitated by the advanced capabilities of heterogeneous GNNs to model physical proximity. We found that our analysis of the functioning genes aligns with previous findings, proving our model’s effectiveness as a potential method for further analyze communication mechanisms.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>