GENtervention: Gene expression changes associated with lifespan-extending interventions

GENtervention is a platform for investigation of associations between longevity and hepatic expression of individual genes in mice.

To cite the use of this app in your publications, please reference:
Tyshkovskiy A, Bozaykut P, Borodinova AA, Gerashchenko MV, Ables GP, Garratt M, Khaitovich P, Clish CB, Miller RA, Gladyshev VN (2019). Identification and Application of Gene Expression Signatures Associated with Lifespan Extension. Cell Metabolism 30, 1-21. doi: 10.1016/j.cmet.2019.06.018.     

GENtervention includes three instruments for visualization of longevity associations:

1. Summary barplot

The first instrument, Summary barplot, allows to investigate general association between hepatic expression of individual genes and lifespan extension. For every gene of interest, it demonstrates aggregated expression response (logFC) to each healthspan- and lifespan-extending intervention. Red asterisk denotes interventions, which significantly change the expression of certain gene.

The following options are available for this tool:

• Randomize by: defines the random term, which is used during calculation of aggregated logFC and p-value. Random term combines related datasets into certain groups and favours the association if it is supported by different groups. Control group (default) aggregates intervention responses if they are compared to the same set of control samples. Source combines datasets if they belong to the same source of data (usually the same Pubmed ID).

• Only significant longevity effect: defines if gene expression should be aggregated across all interventions and experimental settings (No), or only across interventions and experimental conditions, which have been shown to produce statistically significant extension of lifespan in Mus musculus model (Yes, default).

• Adjusted p-value threshold (default: 0.05): sets threshold for Benjamini-Hochberg adjusted p-value corresponding to the null hypothesis that certain gene is neither up- nor downregulated in response to particular intervention (its logFC is equal to 0). P-value is calculated using mixed-effect model for interventions supported by several data sources and is estimated using edgeR model for interventions supported only by RNAseq (e.g., Protandim and 17-alpha-estradiol). Red asterisk denotes interventions with adjusted p-value < current threshold.     
    

2. Common signatures

The second instrument, Common signatures, demonstrates the range of expression change for a particular gene in response to different lifespan- and healthspan-extending interventions across individual datasets, where the gene has been detected. Thus, it allows to estimate if the gene of interest is commonly up- or downregulated after introduction of all or certain interventions. Each dot corresponds to a single dataset and represents mean logFC of the gene in response to particular treatment.

The following options are available for this tool:

• Type of intervention: defines if plot should include datasets corresponding to All interventions (default) or only certain set of treatments: Growth hormone deficiency interventions (Ames, Snell dwarf mice, Little mice and GHRKO), Caloric restriction, Rapamycin, Ames dwarf mice, Little mice or Acarbose.

• Only significant longevity effect: defines if plot should include datasets corresponding to all treatments (No) or only interventions and experimental conditions, which have been shown to produce statistically significant extension of lifespan in Mus musculus model (Yes, default).

• Sex: defines if datasets corresponding to Both sexes (default), Only males or Only females should be included.

• Color dataset by: defines if individual points should be colored based on Type of intervention (default), Source (Pubmed ID) or Sex.     
    

3. Association with the size of effect

Finally, the third instrument, Association with the size of effect, allows to discover positive and negative associations between hepatic expression change of individual gene and the size of lifespan extension effect, defined by different quantitative metrics. Each dot corresponds to a single dataset and represents mean logFC of the gene in response to particular treatment.

The following options are available for this tool:

• Lifespan metric: defines the quantitative metric shown on the x-axis. Available estimates for the size of lifespan extension effect include change of Maximum lifespan (default, logarithm of ratio between average lifespans of 10% most long-lived animals in the cohort), Median lifespan (logarithm of ratio between median lifespans) and logarithm of Median hazard ratio. All values are obtained from published survival data corresponding to Mus musculus model.

• Only significant longevity effect: defines if plot should include datasets corresponding to all treatments (No) or only interventions and experimental conditions, which have been shown to produce statistically significant extension of lifespan in Mus musculus model (Yes, default).

• Sex: defines if datasets corresponding to Both sexes (default), Only males or Only females should be included.

• Color dataset by: defines if individual points should be colored based on Type of intervention (default), Source (Pubmed ID) or Sex.