General content of the DGRPool web tool.

A. Pubmed search on “Drosophila DGRP” terms unveiled 155 results from 2012 to 2024 (search made in July 2024). B. Sex of the DGRP lines used across all 135 studies (left) and 43 curated studies (right), for each phenotype. C. Number of studies per phenotype category. Studies can be assigned to multiple categories. D. Number of phenotypes per study and per sex. Studies without attached phenotypes were not plotted. Of note, a given phenotype can be measured for different sexes and thus counted multiple times.

Comparison of the two currently available web portals organizing DGRP phenotyping data.

This table compares different features available in DGRPool, with DGRP2, the current main resource for DGRP data. It separates the features into 1) Data, which summarizes the available phenotyping data, 2) Tools, which lists the available tools and options, mainly GWAS, PheWAS and phenotype correlation, 3) Web, which describes the website itself, and 4) Additional features, that are available in DGRPool, such as the curation system, the possibility to publish new studies and the interactive plots. Of note, the 838 phenotypes are counted regardless of the associated sex (M, F and/or NA), while the “sex-specific” value is calculated by counting the same phenotype separately for each available sex.

Within- and cross-study phenotype correlations.

A. Spearman’s correlation of all phenotypes available in the 43 curated studies. Of note, we separately computed the phenotype correlations when data per sex were available (M, F or NA), and we restricted the computation to quantitative (non-categorical) phenotypes. Phenotypes are grouped by study (colored box at the bottom of the plot). B. Absolute value of the Spearman’s correlation of pairs of phenotypes that originated from the same study (within-study) and those that originated from two different studies (cross-study). Of note, displayed values are median. Mean values are 0.099 for cross-study, and 0.259 for within-study. Again, we restricted the calculation to the 43 curated studies. C. Correlation of two longevity phenotypes from the same study (Arya, et al, 2010)34, revealing a strong correlation between Female (F) and Male (M) longevity. D. Correlation of two phenotypes from different studies: mean lifespan (Durham et al, 2014)31 and mean longevity (Arya et al, 2010)34. Of note, both the C and D plots were generated using the “phenotype correlation” tool in DGRPool.

Phenotype correlations contribute new biological insights.

A. Correlation of mean femur length (Grubbs et al., 2013)39 vs. mean head width (Vonesch et al., 2016)28 showing the significant cross-study association of organismal size traits. B. Correlation of remating proportion (Chow et al., 2013)41 vs. mean fecundity (Durham et al., 2014)31. C. The 33 phenotypes correlated with longevity (Arya et al, 2010)34 at a 5% FDR threshold (Spearman’s correlation), revealing three main groups of phenotypes: lifespan phenotypes (middle rows), other correlated phenotypes (bottom rows) and anti-correlated phenotypes (top rows). Of note, both the A and B plots were generated using the “phenotype correlation” tool in DGRPool.

Overview of GWAS results across phenotypes and one case study.

A. Distribution of the number of significant variants after a GWAS, for each phenotype available in DGRPool. Of note, all values >50 have been set to 50, for easier visualization. B. For each variant, we plotted the number of times it was significantly associated with a phenotype (y-axis = number of occurrences). It is worth noting that we chose a Manhattan plot for representing this information, but this is not a “real” GWAS Manhattan plot. C. Case study on survival to azinphos-methyl exposure (Battlay et al., 2016)30, here to a 0.25 µg/ml dose. This plot was extracted from the phenotype’s page on DGRPool at https://dgrpool.epfl.ch/phenotypes/20. D. Manhattan plot (taken from DGRPool’s result page https://dgrpool.epfl.ch/phenotypes/20/gwas_analysis) showing the association of variants to “survival at 0.25 µg/ml dose” phenotype. E. Boxplot (taken from DGRPool’s result page https://dgrpool.epfl.ch/phenotypes/20/gwas_analysis), showing the effect of the top variant, 2R:8072884, which is a long insertion.

Analysis of extremeness among DGRP lines across 40 phenotypes.

A. Fraction of extremeness of a given DGRP line. DGRP lines are assigned as ‘extreme’ in a phenotype when they are in the top or bottom 15% of the phenotypic spectrum. Phenotypes were selected based on the curated studies which had the following categories assigned to them: Life history traits, Immunity, Toxicity, Resistance, Fecundity, Aging. DGRP lines were included if they had at least 50 phenotypic measures. B. Scatter plot for the fraction of extremeness of DGRP lines. On the x-axis, the fraction of extremeness is plotted for females, whereas males are plotted on the y-axis. C. Most extreme and moderate DGRP lines per sex. On the x-axis, the adjusted fraction of extremeness is provided. Individual fractions of extremeness per phenotype were retrieved for each DGRP line. The fraction was adjusted by 1 minus the fraction of extremeness if the fraction of extremeness was above 0.5. Because extremeness can range from 0 to 0.15 or 0.85 to 1, we adjusted the fraction of extremeness for plotting purposes. DGRP lines with a low adjusted fraction of extremeness are therefore more extreme, whereas a high adjusted fraction of extremeness is representative of more moderate DGRP lines. D. Extreme and moderate DGRP line pairings. On the x-axis, the adjusted fraction of extremeness is provided. Extreme and moderate line pairings were retrieved by searching for DGRP lines for which the fraction of extremeness between females and males was not greater than 0.05 while still having the highest and lowest average fraction of extremeness (across sex). E. Looking at phenotypes from Figure 2D marked as longevity/lifespan, for DGRP lines which are in the top 5 of fraction of extremeness for each respective sex, including DGRP_852 and DGRP_042 (red shades) from 5D. We specifically highlight DGRP_757, DGRP_765 in blue shades to show that they are across multiple studies in the lower end of the lifespan as is expected given that the lifespan trait is robust across studies. Similarly, DGRP_320 shows a trend in which it displays above average lifespan. Other extreme DGRP lines which were in each respective top 5 are displayed in gray.