Facial_masculinity_MHC

Data and scripts for "Facial masculinity does not appear to be a condition-dependent male ornament and does not reflect MHC heterozygosity in humans"

If you use any of the data or scripts for your own research, please cite the biorxiv paper: \

Zaidi AA*, White JD*, Mattern BC, Liebowitz CR, Puts DA, Claes P, Shriver MD. 2019. Facial masculinity does not appear to be a condition-dependent male ornament and does not reflect MHC heterozygosity in humans. Proceedings of the National Acadedmy of Sciences USA 116:1633–1638. Available from: http://www.ncbi.nlm.nih.gov/pubmed/30647112

Details of folders:

Scripts

1. calc_masc_03262018.R : Script used to generate facial masculinity measures from 3D coorsinates data
2. summary_stats_03262018.R : Script used to calculate summary statistics (e.g. degree of sexual dimorphism, cohen's D, and Levene's test of equal variances) on facial masculinity
3. masc_height_association_03262018.R : Script used to test for associations between facial masculinity and height (Hypothesis 1).
4. height_hla_association_02272018.R : Script used to test for association between height and MHC heterozygosity (Hypothesis 2).
5. masc_hla_association_03262018.R : Script used to test for association between facial masculinity and MHC heterozygosity (Hypothesis 3).
6. Plot1Face.R: Script used to visualize high-dimensional facial masculinity scores, as well as quasi-landmark based test statistics (cohen's D, beta coefficients, p-values etc.) as facial heatmaps.
7. GeneticAnalysisNotes.pdf: A pdf file that walks through all of the processing of genetic data to calculate heterozygosity for this paper. Included are code blocks, log files, explanatory text, and images.

Dataset

1. Euro_demographic_03292018.dat
   Dataframe containing demographic data for all 1,233 individuals used in the study. Columns are in the following order:
   ... IID: Unique identifier for individual
   ... phet_genome: Heterozygosity across genome-wide (LD-pruned) SNPs
   ... phet_hla: Heterozygosity across MHC locus
   ... Sex
   ... Age
   ... Height: In centimeters
   ... Weight: In Kg
   ... gPC1-4: genetic PC scores\

2. euro_1233_masc_het_03292018.dat
   Dataframe containing all the columns from Euro_demographic_03292018.dat and two additional columns:
   a) avg.masc: Average facial masculinity calculated across all 7,150 QLs for each person
   b) avg.masc.unit: Average facial masculinity scaled by the Euclidean distance between female and male consensus faces. In other words, this is average facial masculinity divided by the average facial masculinity of the male consensus face.

3. qlmasc_1233_noprop_03292018.txt.zip
   Zipped text file containing high-dimensional facial masculinity scores (dimensions: 7,150 x 1,233) i.e. 3D facial masculinity calculated for 7,150 quasi-landmarks (rows) for each of 1,233 individuals (columns). Referred to in the paper as FM_QL. No headers or indices included. The order of the individuals is the same as the order of individuals in Euro_demographic_03292018.dat file. The overall facial masculinity scores were calculated by averaging FM_QL across the 7,150 QLs. The FM_QL scores can be used in further analyses or visualized using the Plot1Face.R function. Take a look at the tutorial for more details.

4. Refscan.obj
   .OBJ file containing the 3D facial template used to visualize facial heatmaps etc.

Results

Summary_dat

1. ql_gsd_03292018.txt
   ... dataframe of length 7,150 x 1. Each row is the absolute degree of sexual dimorphism for each quasi-landmark

2. ql_sex_cohenD_03292018.txt
   ... dataframe of length 7,150 x 1. Each row is the Cohen's D of sexual dimorphism for each quasi-landmark

Masc_v_height

Linear model results: facial masculinity ~ height + covariates

1. lm_overallmasc_height_results_03292018
   ... Results of linear model where overall facial masculinity was used as a response

2. qlmasc_height_b_noprop_sd_04142018.txt
   ... Beta coefficients of linear model where FM_QL was used individually as a response

3. qlmasc_height_p_noprop_sd_04142018.txt
   ... p-value for beta coefficients

4. qlmasc_height_q_noprop_sd_04142018.txt
   ... indicator variable to indicate if p-value passes Bonferroni correction ...... 0 if p-value <0.05/7,150 ...... 1 if p-value >= 0.05/7,150

5. qlmasc_h_bysex_noprop_sd_04142018.txt
   ... slope_female: beta coefficients for FM_QL - females only ... slope_male: beta coefficients for FM_QL - males only ... diff: slope_male - slope_female ... z: Z-score for difference in slope ... p.value: Two-sided p-value for Z-score ... q.value: indicator variable indicating if p-value passes Bonferroni correction

Tutorial_visualizing_facial_heatmaps

The Rmarkdown file (.rmd) demonstrates how the FM_ql data, as well as other ql-specific statistics can be viewed on a face. A rendered .html file is provided, which should be downloaded onto your computer first and then opened in a browser. The used of plotly allows the 3D faces to be rotated and scaled.

The screenshots are provided to illustrate what the tutorial looks like.