MBI Seminar Series - Amir Asiaee | Mathematical Biosciences Institute

April 15, 2021

10:00AM - 11:00AM

Virtual Zoom Seminar

Add to Calendar 2021-04-15 10:00:00 2021-04-15 11:00:00 MBI Seminar Series - Amir Asiaee Title: Identification and comparison of genes differentially regulated by transcription factors and miRNAs Abstract: BACKGROUND: Transcription factors (TF) and microRNAs (miR) play critical roles in gene regulation. To better understand which genes are being regulated by TFs and/or miRs, we studied over 8000 samples from 32 cancer types collected from The Cancer Genome Atlas (TCGA). METHODS: We started by clustering the TFs and miRs using Thresher to reduce the number of features. Thresher combines PCA, outlier filtering, and von Mises-Fisher mixture model to cluster features. Thresher reduces the number of TFs and miRs from 486 and 470 to 30 and 21 scores, respectively. We use scores to predict the expression of ~20000 genes by three types of linear models: (i) TF, (ii) miR, and (iii) TF plus miR. We select genes present at the tails of R2 distribution for gene enrichment analysis using ToppGene. RESULTS: TFs and miRs clusters determined by Thresher are biologically interpretable and are statistically associated with specific cancer types. Moreover, the 51 scores contain sufficient information to distinguish between most types of cancer. We found that on average (a) the 21 miR scores explain 30%, (b) the 30 TF scores explain 47%, and (c) the joint model explains 50% of transcriptome variation. The poorly explained genes are enriched for olfactory receptor processes, sensory perception, and nervous system processing which are necessary to receive and interpret signals from outside the organism. In contrast, highly explained genes are characterized by genes translating to proteins essential for transport, plasma membrane, or metabolic processes that are heavily regulated inside the cell. SIGNIFICANCE: We observe that TFs, miRs, or both sufficiently explain the expression of many genes. Other genetic regulatory elements such as methylation may explain the poorly explained tail of the R2 distribution. In the future, we want to map the causal-effect relationship between regulatory elements and transcriptome. Zoom information: To join the seminar by zoom, please use the following link: https://osu.zoom.us/j/93066786961?pwd=aGxpQitJUmNLNlRqSi9naXBmWWx4dz09 Video: https://osu.box.com/s/9ru6pyrnz3h0yoh907cg1x5knbdbh306 Virtual Zoom Seminar OSU ASC Drupal 8 ascwebservices@osu.edu America/New_York public

Add to Calendar 2021-04-15 10:00:00 2021-04-15 11:00:00 MBI Seminar Series - Amir Asiaee Title: Identification and comparison of genes differentially regulated by transcription factors and miRNAs Abstract: BACKGROUND: Transcription factors (TF) and microRNAs (miR) play critical roles in gene regulation. To better understand which genes are being regulated by TFs and/or miRs, we studied over 8000 samples from 32 cancer types collected from The Cancer Genome Atlas (TCGA).  METHODS: We started by clustering the TFs and miRs using Thresher to reduce the number of features. Thresher combines PCA, outlier filtering, and von Mises-Fisher mixture model to cluster features. Thresher reduces the number of TFs and miRs from 486 and 470 to 30 and 21 scores, respectively. We use scores to predict the expression of ~20000 genes by three types of linear models: (i) TF, (ii) miR, and (iii) TF plus miR. We select genes present at the tails of R2 distribution for gene enrichment analysis using ToppGene.  RESULTS: TFs and miRs clusters determined by Thresher are biologically interpretable and are statistically associated with specific cancer types. Moreover, the 51 scores contain sufficient information to distinguish between most types of cancer. We found that on average (a) the 21 miR scores explain 30%, (b) the 30 TF scores explain 47%, and (c) the joint model explains 50% of transcriptome variation. The poorly explained genes are enriched for olfactory receptor processes, sensory perception, and nervous system processing which are necessary to receive and interpret signals from outside the organism. In contrast, highly explained genes are characterized by genes translating to proteins essential for transport, plasma membrane, or metabolic processes that are heavily regulated inside the cell.  SIGNIFICANCE: We observe that TFs, miRs, or both sufficiently explain the expression of many genes. Other genetic regulatory elements such as methylation may explain the poorly explained tail of the R2 distribution. In the future, we want to map the causal-effect relationship between regulatory elements and transcriptome.   Zoom information: To join the seminar by zoom, please use the following link:  https://osu.zoom.us/j/93066786961?pwd=aGxpQitJUmNLNlRqSi9naXBmWWx4dz09 Video:  https://osu.box.com/s/9ru6pyrnz3h0yoh907cg1x5knbdbh306   Virtual Zoom Seminar Mathematical Biosciences Institute mbi-webmaster@osu.edu America/New_York public

Title: Identification and comparison of genes differentially regulated by transcription factors and miRNAs

Abstract:

BACKGROUND: Transcription factors (TF) and microRNAs (miR) play critical roles in gene regulation. To better understand which genes are being regulated by TFs and/or miRs, we studied over 8000 samples from 32 cancer types collected from The Cancer Genome Atlas (TCGA).

METHODS: We started by clustering the TFs and miRs using Thresher to reduce the number of features. Thresher combines PCA, outlier filtering, and von Mises-Fisher mixture model to cluster features. Thresher reduces the number of TFs and miRs from 486 and 470 to 30 and 21 scores, respectively. We use scores to predict the expression of ~20000 genes by three types of linear models: (i) TF, (ii) miR, and (iii) TF plus miR. We select genes present at the tails of R² distribution for gene enrichment analysis using ToppGene.

RESULTS: TFs and miRs clusters determined by Thresher are biologically interpretable and are statistically associated with specific cancer types. Moreover, the 51 scores contain sufficient information to distinguish between most types of cancer. We found that on average (a) the 21 miR scores explain 30%, (b) the 30 TF scores explain 47%, and (c) the joint model explains 50% of transcriptome variation. The poorly explained genes are enriched for olfactory receptor processes, sensory perception, and nervous system processing which are necessary to receive and interpret signals from outside the organism. In contrast, highly explained genes are characterized by genes translating to proteins essential for transport, plasma membrane, or metabolic processes that are heavily regulated inside the cell.

SIGNIFICANCE: We observe that TFs, miRs, or both sufficiently explain the expression of many genes. Other genetic regulatory elements such as methylation may explain the poorly explained tail of the R² distribution. In the future, we want to map the causal-effect relationship between regulatory elements and transcriptome.

Zoom information: To join the seminar by zoom, please use the following link:

https://osu.zoom.us/j/93066786961?pwd=aGxpQitJUmNLNlRqSi9naXBmWWx4dz09

Video: https://osu.box.com/s/9ru6pyrnz3h0yoh907cg1x5knbdbh306

Events Filters:

Seminar