Improved Inference of Ecological Interaction Types

Inference of microbial interaction types allows us to understand the growth and development of microbial life forms found on earth. Numerous methods have been proposed to infer the interaction type(s) of microbes in a microbial communities using a population dynamics model. However, due to dynamic behaviour of microbial communities, these methods can result in erroneous inferences. A method proposed by Xiao et al. in 2017 models the dynamic behaviour of microbial community using sample abundance data overcomes many of these issues, but suffers from a high failure rate of inference, lower confidence on inferred interactions and slower execution speed than the existing algorithms. In this thesis, we propose an improved and more efficient and effective approach to infer the microbial interaction types of larger microbial communities (N>10). Our findings demonstrate that our approach is faster, more fault tolerant, more scalable than the state of the art from 2017, and it has the ability to infer microbial interactions with increased confidence.

Name thesis:degree_name

Master of Science (M.Sc.)

Level thesis:degree_level

Masters

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Saskatchewan

Year dc:date.issued

2020

Author dc:creator

• Aziz, Syed Umair

Advisor dc:contributor.advisor

• Stanley, Kevin

Committee members dc:contributor.committeemember

• Kusalik, Tony
• Siciliano, Steven
• Mondal, Debajyoti
• Peak, Derek

• microbial interactions
• unsupervised

Handle dc:identifier.uri

https://hdl.handle.net/10388/13096

OAI identifier oai:identifier

oai:harvest.usask.ca:10388/13096