/trost_phenotypes

trost_phenotypes

BBCH Data

Plant Height Data

FW/DW Ratios

RWC Data

Soil Moisture Sensor Data

Yield Data Analysis

Yield Data from TROST project (2011-2013)

  • yield_data.Rmd

Yield Data from VALDIS project (2014-2015)

  • yield_data_valdis.Rmd
  • use func_get_yield_data to retrieve raw data from Phenotyper DB → yield_data
  • do some modifications on meta information (treatment, genotype)
  • handle NAs, remove dupicate entries, find outlier → yield_data_no_duplicates
  • create subsets for tuber FW, starch content (SC) and starch yield (SY) → examine histograms
  • get subsets for single experiments and remove additional duplicates (e.g. yield_data_67518)
  • extract the commonly used set of lines in 2014:
    • 192 were present in all 4 experiments + AR21 (not in MPI field, but belongs to SP1) → finally 193 lines were considered (lines_2014)
  • extract the commonly used set of lines in 2015:
    • 60 were present in all 5 experiments (common_lines_2015)
    • 3 additional line were present in MPI FGH, MPI field and JKI shelter, but missing in Dethlingen and JKI field (all_lines_2015)
  • create correct yield data table: calculate mean per plant_ID to get rid of replicated measurements (FW, SC) → yield_data_correct
  • replace NA values of SC by ZERO if FW < 0.1kg for the respective plant → this will influence the calculation of mean SY in further analysis
    • if FW < 0.1kg, SY will be zero (52 entries)
    • if FW > 0.1kg and if SC is NA, SY will be NA (9 entries)
  • get information about subpopulations
    • sp_infos → every line is only listed once
    • sp_infos_dup → every line that belongs to two different SPs is duplicated
  • merge data from correct data from 2014 and 2015 with SP information → yield_data_2014_sp_dup and yield_data_2015_sp_dup
  • create subsets for tuber FW, SC and SY → examine histograms
  • calculate SY per experiment with func_starch_yield_feld
    • results in new table for each experiment (e.g. mpi_fgh_2014)
    • calculate coefficient of variation (CV) of SY per line and treatment of each experiment
    • calculate mean of SY per line and treatment of each experiment
  • get subset for control conditions per experiment (e.g. mpi_fgh_2014_control)
    • calculate mean of starch yield in g per plant for each line (only control data!) (e.g. mpi_fgh_2014_control_mean_per_line)
    • calculate overall median of mean values per line (only control data!) → results in one value per experiment (e.g. mpi_fgh_2014_control_median)
    • calculate median of starch yield in g per plant for each line (only control data!) (e.g. mpi_fgh_2014_control_median_per_line)
  • normalize starch yield per experiment: use control median of each experiment to calculate ratio → results in new column e.g. mpi_fgh_2014$normalized_starch_yield_per_plant
  • bind all processed yield data per year → yield_data_final_2014 and yield_data_final_2015
  • the final dataset contains:
    • tuber_FW_kg_per_plot
    • starch_g_per_kg
    • starch_yield_g_per_plant
    • starch_yield_g_per_plot
    • normalized_starch_yield_per_plant
  • get subset for control conditions of final data → yield_data_final_2014_control and yield_data_final_2015_control
    • calculate mean of starch yield per line and experiment (e.g. yield_data_final_2014_control_sy_mean)
    • calculate mean of NORMALIZED starch yield per line and experiment (yield_data_final_2014_control_norm_sy_mean)
  • get subset for stress conditions of final data → yield_data_final_2014_stress and yield_data_final_2015_stress
    • calculate mean of starch yield per line and experiment (e.g. yield_data_final_2014_stress_sy_mean)
  • get subsets for subpopulations
  • create explorative plots for:
    • abolute starch yield ~ treatment
    • normalized starch yield ~ treatment
    • absolute starch yield ~ treatment * plant line (also ordered)
    • normalized starch yield ~ treatment * plant line (also ordered)
    • absolute starch yield ~ treatment * SP
    • normalized starch yield ~ treatment * SP
  • calculate ANOVA for starch yield ~ treatment * plant line
  • calculate stress index (SI) = 1 - (mean(SY_drought) / mean(SY_control)
    • all_SI_list_2014
    • all_SI_list_2015
    • or per experiment, e.g. mpi_fgh_2014_si
  • calculate Relative Starch Yield (RelSY), e.g. mpi_fgh_2014_relSY
    • single value per replicate of drought stress (mpi_fgh_2014_relSY)
    • median per line (mpi_fgh_2014_relSY_median)
  • calculate DRYM
    • single value per replicate of drought stress (mpi_fgh_2014_drym)
    • median per line (mpi_fgh_2014_drym_median)
  • combine DRYM median per line per experiment (drym_experiments_2014 and drym_experiments_2015)
    • compare DRYM of SPs by t-test
    • get subset per experiment (drym_mpi_fgh_2014)
    • plot DRYM per experiment and SP (also as single lines, grouped by SP)
    • ANOVA of DRYM
    • plot DRYM versus normalized SY → yield penalty!
    • pairs plot of DRYM and cor.test (between experiments)
  • plot of starch yield vs. starch content
  • calculate Stress Sensitivity Index (SSI)
    • High value of SSI corresponds to sensitive genotype
    • Low value of SSI corresponds to tolerant genotype
  • plot SSI versus DRYM