UKB phenotype data formatting
by [Baptiste Couvy-Duchesne] - 23 April 2020 -
1 Open data
We have created a csv file using the R package ukbtools which contains all phenotypes available as part of our application, but limited to the participants for which we have brain images
This reduces massively the size of the data and allows quicker handler with R
We use the column names generated with ukbtools but do not use data completely recoded as we want to keep numeric values in the variables, even for discrete variables
This is because OSCA does not accept factors, and requires discrete phenotypes to be coded with numbers
2 Create new variables or reformat into more meaninful variables
Note that we have found different versions of ukbtools (typically between on discovery and replication analyses) to behave slightly differently, in particular when setting new variable names
This may explain some problems you may experience when running our code (check your variable names, date format, and labels of discrete variables)
# Age at MRI
TT$Age_MRI<-NA
TT$Date_birth<-paste("01", TT$month_of_birth_f52_0_0, TT$year_of_birth_f34_0_0, sep = "-")
TT$Date_birth<-as.Date(TT$Date_birth, "%d-%m-%Y")
TT$date_of_attending_assessment_centre_f53_2_0<-as.Date(TT$date_of_attending_assessment_centre_f53_2_0, "%Y-%m-%d")
TT$Age_MRI<-as.numeric(TT$date_of_attending_assessment_centre_f53_2_0-TT$Date_birth)/365.25
hist(TT$Age_MRI, breaks = 500)
length(which(is.na(TT$Age_MRI))) # No missing value
# For example: Sex variable changed name when updating ukbtools - sexuses_datacoding_9_f31_0_0 coded 0-1 became sex_f31_0_0 coded Female-Male
# We duplicated the sex variable in sexuses_datacoding_9_f31_0_0 so the code did not have to be changed in whole analysis
TT$sexuses_datacoding_9_f31_0_0=TT$sex_f31_0_0
# Create some cognition variables
TT$percent_correct_digit_matches_0_0<-TT$number_of_symbol_digit_matches_made_correctly_f20159_0_0/TT$number_of_symbol_digit_matches_attempted_f20195_0_0
TT$percent_correct_digit_matches_0_0_binary=ifelse(TT$percent_correct_digit_matches_0_0==1, 1, 0)
# Number of diagnoses - 15045 participants extracted from the mental health section (10743)
# Meaning all missing values are negative answers (i.e. none of the above)
TT$Number_mental_health_problems_diagnoses_f20544=rowSums(!is.na(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", 1:16)]))
# Extract specific diagnoses
TT$SocialAnxietySocialPhobia=0
TT$Schizophrenia=0
TT$PsychosisNotSchizophrenia=0
TT$PersonalityDisorder=0
TT$Phobia=0
TT$PanicAttacks=0
TT$OCD=0
TT$ManiaBipolar=0
TT$Depression=0
TT$Bulimia=0
TT$BingeEating=0
TT$Autism=0
TT$AnxietyGAD=0
TT$Anorexia=0
TT$Agoraphobia=0
TT$ADHD=0
for (iii in 1:16){
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==1 ),"SocialAnxietySocialPhobia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==2 ),"Schizophrenia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==3 ),"PsychosisNotSchizophrenia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==4 ),"PersonalityDisorder"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==5 ),"Phobia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==6 ),"PanicAttacks"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==7 ),"OCD"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==10 ),"ManiaBipolar"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==11 ),"Depression"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==12 ),"Bulimia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==13 ),"BingeEating"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==14),"Autism"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==15 ),"AnxietyGAD"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==16 ),"Anorexia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==17 ),"Agoraphobia"]=1
TT[which(TT[,paste0("mental_health_problems_ever_diagnosed_by_a_professionaluses_datacoding_1401_f20544_0_", iii)]==18 ),"ADHD"]=1
}
# Duplicate a few more variables
# tinnitus
TT$tinnitususes_datacoding_100635_f4803_2_0_binary=TT$tinnitususes_datacoding_100635_f4803_2_0
TT$tinnitususes_datacoding_100635_f4803_2_0_present=TT$tinnitususes_datacoding_100635_f4803_2_0
# Prospective memory
TT$prospective_memory_result_3points_f20018_2_0=TT$prospective_memory_result_f20018_2_0
TT$prospective_memory_result_3points_f20018_2_0[which(TT$prospective_memory_result_3points_f20018_2_0==1)]=4
TT$prospective_memory_result_3points_f20018_2_0=TT$prospective_memory_result_3points_f20018_2_0/2
table(TT$prospective_memory_result_3points_f20018_2_0)3 List of variables, categories and labels
Throughout the analyses, we use a few files that contain the variable names, categories and labels
They facilitate plotting and looping on variables in the analyses
They are available in the GitHub repository
The files we share contain more variables than what is presented in the manuscript because it already includes variables we will use in a follow up manuscript
Contact us, if you are interested by results not published yet
4 Format phenotyes and exclude missing values
winsoriseNA<-function(var){
lb<-mean(TT[,var], na.rm = T)-4*sd(TT[,var], na.rm = T)
ub<-mean(TT[,var], na.rm = T)+4*sd(TT[,var], na.rm = T)
TT[which(TT[,var]>ub),var]<-NA
TT[which(TT[,var]<lb),var]<-NA
return(TT[,var])
}4.1 Demographics
UKBvar$V4[which(UKBvar$V2=="Demographics")]
TT$average_total_household_income_before_tax_f738_2_0[which(TT$average_total_household_income_before_tax_f738_2_0<0)]<-NA
TT$age_completed_full_time_education_f845_2_0[which(TT$age_completed_full_time_education_f845_2_0<0)]<-NA
TT$number_in_household_f709_2_0[which(TT$number_in_household_f709_2_0<0)]<-NA
TT$number_in_household_f709_2_0[which(TT$number_in_household_f709_2_0>5)]<-6
table(TT$number_in_household_f709_2_0)4.2 Blood assay
We excluded outlying values +-4SD because we observed large tails in distributions
4.3 Brain phenotypes
ICV and positive control variables
4.4 Cognition
UKBvar$V4[which(UKBvar$V2=="Cognition")]
prop.table(table(TT$prospective_memory_result_f20018_2_0)) # recode into - recalled vs. non recalled
TT$prospective_memory_result_f20018_2_0[which(TT$prospective_memory_result_f20018_2_0==2)]=1
TT$prospective_memory_result_f20018_2_0=(TT$prospective_memory_result_f20018_2_0 -1 )*(-1)
TT$maximum_digits_remembered_correctly_f20240_0_0[which(TT$maximum_digits_remembered_correctly_f20240_0_0<0)]<-NA
TT$maximum_digits_remembered_correctly_f4282_0_0[which(TT$maximum_digits_remembered_correctly_f4282_0_0<0)]<-NA
TT$total_errors_traversing_numeric_path_trail_1_f20247_0_0=winsoriseNA("total_errors_traversing_numeric_path_trail_1_f20247_0_0")
TT$total_errors_traversing_alphanumeric_path_trail_2_f20248_0_0=winsoriseNA("total_errors_traversing_alphanumeric_path_trail_2_f20248_0_0")4.5 Diet
4.6 Miscelanious
UKBvar$V4[which(UKBvar$V2=="Other")]
TT$diabetes_diagnosed_by_doctoruses_datacoding_100349_f2443_2_0[which(TT$diabetes_diagnosed_by_doctoruses_datacoding_100349_f2443_2_0<0)]<-NA
TT$overall_health_rating_f2178_2_0[which(TT$overall_health_rating_f2178_2_0 %in% c(-1,-3))]<-NA
# reverse code - left handed and anbidextrueous - then replace phenotype
TT$handedness_chiralitylateralityuses_datacoding_100430=TT$handedness_chiralitylaterality_f1707_2_0
TT$handedness_chiralitylateralityuses_datacoding_100430[which(TT$handedness_chiralitylaterality_f1707_2_0 %in% c(-1,-3))]<-NA
TT$handedness_chiralitylateralityuses_datacoding_100430[which(TT$handedness_chiralitylaterality_f1707_2_0 ==2)]<-3
TT$handedness_chiralitylateralityuses_datacoding_100430[which(TT$handedness_chiralitylaterality_f1707_2_0 ==3)]<-2
TT$handedness_chiralitylaterality_f1707_2_0=TT$handedness_chiralitylateralityuses_datacoding_100430
TT$sleep_duration_f1160_2_0[which(TT$sleep_duration_f1160_2_0 %in% c(-1,-3))]<-NA
TT$sleeplessness_insomnia_f1200_2_0[which(TT$sleeplessness_insomnia_f1200_2_0 %in% c(-1,-3))]<-NA
TT$time_spent_watching_television_tv_f1070_2_0[which(TT$time_spent_watching_television_tv_f1070_2_0 %in% c(-1,-3))]<-NA
TT$time_spent_watching_television_tv_f1070_2_0[which(TT$time_spent_watching_television_tv_f1070_2_0==-10)]<-0
TT$part_of_a_multiple_birth_f1777_2_0[which(TT$part_of_a_multiple_birth_f1777_2_0<0)]<-NA
TT$general_happinessuses_datacoding_537_f20458_0_0[which(TT$general_happinessuses_datacoding_537_f20458_0_0<0)]<-NA
TT$general_happiness_with_own_healthuses_datacoding_537_f20459_0_0[which(TT$general_happiness_with_own_healthuses_datacoding_537_f20459_0_0<0)]<-NA
TT$headaches_for_3_monthsuses_datacoding_100349_f3799_2_0[which(TT$headaches_for_3_monthsuses_datacoding_100349_f3799_2_0<0)]<-NA
TT$hearing_difficultyproblemsuses_datacoding_100631_f2247_2_0[which(TT$hearing_difficultyproblemsuses_datacoding_100631_f2247_2_0<0)]<-NA
TT$hearing_difficultyproblemsuses_datacoding_100631_f2247_2_0[which(TT$hearing_difficultyproblemsuses_datacoding_100631_f2247_2_0==99)]<-1
TT$hearing_difficultyproblems_with_background_noiseuses_datacoding_100349_f2257_2_0[which(TT$hearing_difficultyproblems_with_background_noiseuses_datacoding_100349_f2257_2_0<0)]<-NA
TT$hearing_aid_useruses_datacoding_100352_f3393_2_0[which(TT$hearing_aid_useruses_datacoding_100352_f3393_2_0<0)]<-NA
TT$tinnitususes_datacoding_100635_f4803_2_0_binary[which(TT$tinnitususes_datacoding_100635_f4803_2_0_binary<0)]=NA
TT$tinnitususes_datacoding_100635_f4803_2_0_binary[which(TT$tinnitususes_datacoding_100635_f4803_2_0_binary>0)]=1
TT$tinnitususes_datacoding_100635_f4803_2_0_present[which(TT$tinnitususes_datacoding_100635_f4803_2_0_present<0)]=NA
TT$tinnitususes_datacoding_100635_f4803_2_0_present[which(TT$tinnitususes_datacoding_100635_f4803_2_0_present==14)]=0
TT$tinnitususes_datacoding_100635_f4803_2_0_present[which(TT$tinnitususes_datacoding_100635_f4803_2_0_present==13)]=1
TT$tinnitususes_datacoding_100635_f4803_2_0_present[which(TT$tinnitususes_datacoding_100635_f4803_2_0_present==12)]=2
TT$tinnitususes_datacoding_100635_f4803_2_0_present[which(TT$tinnitususes_datacoding_100635_f4803_2_0_present==11)]=3
TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0[which(TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0<0)]=NA
TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0[which(TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0==4)]=0
TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0[which(TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0==13)]=1
TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0[which(TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0==12)]=2
TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0[which(TT$tinnitus_severitynuisanceuses_datacoding_100636_f4814_2_0==11)]=3
TT$wears_glasses_or_contact_lensesuses_datacoding_100352_f2207_2_0[which(TT$wears_glasses_or_contact_lensesuses_datacoding_100352_f2207_2_0<0)]=NA
TT$time_spend_outdoors_in_summer_f1050_2_0[which(TT$time_spend_outdoors_in_summer_f1050_2_0 %in% c(-1,-3))]<-NA
TT$time_spend_outdoors_in_summer_f1050_2_0[which(TT$time_spend_outdoors_in_summer_f1050_2_0 ==-10)]<-0
TT$time_spent_outdoors_in_winter_f1060_2_0[which(TT$time_spent_outdoors_in_winter_f1060_2_0 %in% c(-1,-3))]<-NA
TT$time_spent_outdoors_in_winter_f1060_2_0[which(TT$time_spent_outdoors_in_winter_f1060_2_0 ==-10)]<-0
TT$time_spent_using_computer_f1080_2_0[which(TT$time_spent_using_computer_f1080_2_0 %in% c(-1,-3))]<-NA
TT$time_spent_using_computer_f1080_2_0[which(TT$time_spent_using_computer_f1080_2_0 ==-10)]<-0
TT$time_spent_driving_f1090_2_0[which(TT$time_spent_driving_f1090_2_0 %in% c(-1,-3))]<-NA
TT$time_spent_driving_f1090_2_0[which(TT$time_spent_driving_f1090_2_0 ==-10)]<-0
TT$morningevening_person_chronotype_f1180_2_0[which(TT$morningevening_person_chronotype_f1180_2_0<0)]=NA
TT$daytime_dozing_sleeping_narcolepsy_f1220_2_0[which(TT$daytime_dozing_sleeping_narcolepsy_f1220_2_0<0)]=NA
TT$breastfed_as_a_baby_f1677_2_0[which(TT$breastfed_as_a_baby_f1677_2_0<0)]=NA
TT$maternal_smoking_around_birth_f1787_2_0[which(TT$maternal_smoking_around_birth_f1787_2_0<0)]=NA
TT$started_insulin_within_one_year_diagnosis_of_diabetesuses_datacoding_100349_f2986_2_0[which(TT$started_insulin_within_one_year_diagnosis_of_diabetesuses_datacoding_100349_f2986_2_0<0)]=NA4.7 Physical test
UKBvar$V4[which(UKBvar$V2=="Physical \n test")]
TT$number_of_daysweek_walked_10_minutes_f864_2_0[which(TT$number_of_daysweek_walked_10_minutes_f864_2_0<0)]<-NA
TT$number_of_daysweek_of_moderate_physical_activity_10_minutes_f884_2_0[which(TT$number_of_daysweek_of_moderate_physical_activity_10_minutes_f884_2_0<0)]<-NA
TT$number_of_daysweek_of_vigorous_physical_activity_10_minutes_f904_2_0[which(TT$number_of_daysweek_of_vigorous_physical_activity_10_minutes_f904_2_0<0)]<-NA
TT$overall_acceleration_average_f90012_0_0=winsoriseNA("overall_acceleration_average_f90012_0_0")
TT$heel_bone_mineral_density_bmd_tscore_automated_f78_0_0=winsoriseNA("heel_bone_mineral_density_bmd_tscore_automated_f78_0_0")
TT$heel_bone_mineral_density_bmd_f3148_0_0=winsoriseNA("heel_bone_mineral_density_bmd_f3148_0_0")4.8 Psychiatry
UKBvar$V4[which(UKBvar$V2=="Psychiatry")]
TT$seen_a_psychiatrist_for_nerves_anxiety_tension_or_depression_f2100_2_0[which(TT$seen_a_psychiatrist_for_nerves_anxiety_tension_or_depression_f2100_2_0 %in% c(-1,-3))]<-NA
TT$seen_doctor_gp_for_nerves_anxiety_tension_or_depression_f2090_2_0[which(TT$seen_doctor_gp_for_nerves_anxiety_tension_or_depression_f2090_2_0 %in% c(-1,-3))]<-NA
TT$impact_on_normal_roles_during_worst_period_of_anxietyuses_datacoding_510_f20418_0_0[which(TT$impact_on_normal_roles_during_worst_period_of_anxietyuses_datacoding_510_f20418_0_0<0)]<-NA
TT$impact_on_normal_roles_during_worst_period_of_depressionuses_datacoding_510_f20440_0_0[which(TT$impact_on_normal_roles_during_worst_period_of_depressionuses_datacoding_510_f20440_0_0<0)]<-NA
TT$ever_felt_worried_tense_or_anxious_for_most_of_a_month_or_longeruses_datacoding_502_f20421_0_0[which(TT$ever_felt_worried_tense_or_anxious_for_most_of_a_month_or_longeruses_datacoding_502_f20421_0_0<0)]<-NA
TT$age_at_first_episode_of_depressionuses_datacoding_513_f20433_0_0[which(TT$age_at_first_episode_of_depressionuses_datacoding_513_f20433_0_0<0)]<-NA
TT$ever_had_prolonged_feelings_of_sadness_or_depressionuses_datacoding_503_f20446_0_0[which(TT$ever_had_prolonged_feelings_of_sadness_or_depressionuses_datacoding_503_f20446_0_0<0)]<-NA
TT$ever_had_prolonged_loss_of_interest_in_normal_activitiesuses_datacoding_503_f20441_0_0[which(TT$ever_had_prolonged_loss_of_interest_in_normal_activitiesuses_datacoding_503_f20441_0_0<0)]<-NA
TT$ever_had_period_of_mania_excitabilityuses_datacoding_502_f20501_0_0[which(TT$ever_had_period_of_mania_excitabilityuses_datacoding_502_f20501_0_0<0)]<-NA
TT$ever_had_period_extreme_irritabilityuses_datacoding_502_f20502_0_0[which(TT$ever_had_period_extreme_irritabilityuses_datacoding_502_f20502_0_0<0)]<-NA
TT$thoughts_of_death_during_worst_depressionuses_datacoding_502_f20437_0_0[which(TT$thoughts_of_death_during_worst_depressionuses_datacoding_502_f20437_0_0<0)]<-NA
TT$lifetime_number_of_depressed_periodsuses_datacoding_511_f20442_0_0[which(TT$lifetime_number_of_depressed_periodsuses_datacoding_511_f20442_0_0<0)]<-NA
TT$lifetime_number_of_depressed_periodsuses_datacoding_511_f20442_0_0[which(TT$lifetime_number_of_depressed_periodsuses_datacoding_511_f20442_0_0>10)]<-10
# Psychosis
TT$ever_believed_in_an_unreal_conspiracy_against_selfuses_datacoding_502_f20468_0_0[which(TT$ever_believed_in_an_unreal_conspiracy_against_selfuses_datacoding_502_f20468_0_0<0)]<-NA
TT$ever_believed_in_unreal_communications_or_signsuses_datacoding_502_f20474_0_0[which(TT$ever_believed_in_unreal_communications_or_signsuses_datacoding_502_f20474_0_0<0)]<-NA
TT$ever_heard_an_unreal_voiceuses_datacoding_502_f20463_0_0[which(TT$ever_heard_an_unreal_voiceuses_datacoding_502_f20463_0_0<0)]<-NA
TT$ever_seen_an_unreal_visionuses_datacoding_502_f20471_0_0[which(TT$ever_seen_an_unreal_visionuses_datacoding_502_f20471_0_0<0)]<-NA
TT$ever_talked_to_a_health_professional_about_unusual_or_psychotic_experiencesuses_datacoding_502_f20477_0_0[which(TT$ever_talked_to_a_health_professional_about_unusual_or_psychotic_experiencesuses_datacoding_502_f20477_0_0<0)]<-NA
# Suicide / self harm
TT$ever_selfharmeduses_datacoding_503_f20480_0_0[which(TT$ever_selfharmeduses_datacoding_503_f20480_0_0<0)]<-NA
TT$ever_thought_that_life_not_worth_livinguses_datacoding_535_f20479_0_0[which(TT$ever_thought_that_life_not_worth_livinguses_datacoding_535_f20479_0_0<0)]<-NA
TT$ever_contemplated_selfharmuses_datacoding_535_f20485_0_0[which(TT$ever_contemplated_selfharmuses_datacoding_535_f20485_0_0<0)]<-NA
# More psychiatry
TT$severity_of_problems_due_to_mania_or_irritabilityuses_datacoding_516_f20493_0_0[which(TT$severity_of_problems_due_to_mania_or_irritabilityuses_datacoding_516_f20493_0_0<0)]<-NA
TT$ever_sought_or_received_professional_help_for_mental_distressuses_datacoding_502_f20499_0_0[which(TT$ever_sought_or_received_professional_help_for_mental_distressuses_datacoding_502_f20499_0_0<0)]<-NA
TT$ever_suffered_mental_distress_preventing_usual_activitiesuses_datacoding_502_f20500_0_0[which(TT$ever_suffered_mental_distress_preventing_usual_activitiesuses_datacoding_502_f20500_0_0<0)]<-NA
# Recent feelings
UKBvar$V4[which(UKBvar$V2=="Recent \n feelings")]
recent<-UKBvar$V4[which(UKBvar$V2=="Recent \n feelings" & !UKBvar$V4 %in% c("recentDepression" , "recentAnxiety" ))]
for (iii in recent){
TT[which(TT[,iii]<0),iii]=NA
}
# Sum of recent feelings: Anx and Dep - recalculate
TT$recentAnxiety<-rowSums(TT[,c("recent_easy_annoyance_or_irritabilityuses_datacoding_504_f20505_0_0", "recent_feelings_of_forebodinguses_datacoding_504_f20512_0_0","recent_inability_to_stop_or_control_worryinguses_datacoding_504_f20509_0_0", "recent_feelings_or_nervousness_or_anxietyuses_datacoding_504_f20506_0_0", "recent_restlessnessuses_datacoding_504_f20516_0_0", "recent_trouble_relaxinguses_datacoding_504_f20515_0_0", "recent_worrying_too_much_about_different_thingsuses_datacoding_504_f20520_0_0")])
TT$recentDepression<-rowSums(TT[,c("recent_changes_in_speedamount_of_moving_or_speakinguses_datacoding_504_f20518_0_0", "recent_feelings_of_depressionuses_datacoding_504_f20510_0_0", "recent_feelings_of_inadequacyuses_datacoding_504_f20507_0_0", "recent_feelings_of_tiredness_or_low_energyuses_datacoding_504_f20519_0_0", "recent_lack_of_interest_or_pleasure_in_doing_thingsuses_datacoding_504_f20514_0_0", "recent_poor_appetite_or_overeatinguses_datacoding_504_f20511_0_0", "recent_thoughts_of_suicide_or_selfharmuses_datacoding_504_f20513_0_0","recent_trouble_concentrating_on_thingsuses_datacoding_504_f20508_0_0", "trouble_falling_or_staying_asleep_or_sleeping_too_muchuses_datacoding_504_f20517_0_0" )])
# Substance use
UKBvar$V4[which(UKBvar$V2=="Substance \n use")]
TT$alcohol_intake_frequency_f1558_2_0[which(TT$alcohol_intake_frequency_f1558_2_0<0)]<-NA
TT$smoking_status_f20116_2_0[which(TT$smoking_status_f20116_2_0==-3)]<-NA
TT$past_tobacco_smoking_f1249_2_0[which(TT$past_tobacco_smoking_f1249_2_0 %in% c(-1,-3))]<-NA
TT$current_tobacco_smoking_f1239_2_0[which(TT$current_tobacco_smoking_f1239_2_0 %in% c(-1,-3))]<-NA
TT$exposure_to_tobacco_smoke_at_homeuses_datacoding_100291_f1269_2_0[which(TT$exposure_to_tobacco_smoke_at_homeuses_datacoding_100291_f1269_2_0<0)]<-NA
TT$exposure_to_tobacco_smoke_outside_homeuses_datacoding_100291_f1279_2_0[which(TT$exposure_to_tobacco_smoke_outside_homeuses_datacoding_100291_f1279_2_0<0)]<-NA
TT$number_of_cigarettes_previously_smoked_dailyuses_datacoding_100353_f2887_2_0[which(TT$number_of_cigarettes_previously_smoked_dailyuses_datacoding_100353_f2887_2_0==-10)]<-0
TT$number_of_cigarettes_previously_smoked_dailyuses_datacoding_100353_f2887_2_0[which(TT$number_of_cigarettes_previously_smoked_dailyuses_datacoding_100353_f2887_2_0==-1)]<-NA
TT$number_of_cigarettes_currently_smoked_daily_current_cigarette_smokersuses_datacoding_100355_f3456_2_0[which(TT$number_of_cigarettes_currently_smoked_daily_current_cigarette_smokersuses_datacoding_100355_f3456_2_0<0)]<-NA
TT$ever_stopped_smoking_for_6_monthsuses_datacoding_100349_f2907_2_0[which(TT$ever_stopped_smoking_for_6_monthsuses_datacoding_100349_f2907_2_0<0)]<-NA
# Addiction
TT$ever_addicted_to_any_substance_or_behaviouruses_datacoding_502_f20401_0_0[which(TT$ever_addicted_to_any_substance_or_behaviouruses_datacoding_502_f20401_0_0<0)]<-NA
TT$ever_addicted_to_alcoholuses_datacoding_502_f20406_0_0[which(TT$ever_addicted_to_alcoholuses_datacoding_502_f20406_0_0<0)]<-NA
TT$ever_addicted_to_a_behaviour_or_miscellanoususes_datacoding_502_f20431_0_0[which(TT$ever_addicted_to_a_behaviour_or_miscellanoususes_datacoding_502_f20431_0_0<0)]<-NA
TT$ever_addicted_to_illicit_or_recreational_drugsuses_datacoding_502_f20456_0_0[which(TT$ever_addicted_to_illicit_or_recreational_drugsuses_datacoding_502_f20456_0_0<0)]<-NA
TT$ever_addicted_to_prescription_or_overthecounter_medicationuses_datacoding_502_f20503_0_0[which(TT$ever_addicted_to_prescription_or_overthecounter_medicationuses_datacoding_502_f20503_0_0<0)]<-NA
TT$amount_of_alcohol_drunk_on_a_typical_drinking_dayuses_datacoding_522_f20403_0_0[which(TT$amount_of_alcohol_drunk_on_a_typical_drinking_dayuses_datacoding_522_f20403_0_0<0)]<-NA
TT$ever_physically_dependent_on_alcoholuses_datacoding_502_f20404_0_0[which(TT$ever_physically_dependent_on_alcoholuses_datacoding_502_f20404_0_0<0)]<-NA
TT$frequency_of_drinking_alcoholuses_datacoding_521_f20414_0_0[which(TT$frequency_of_drinking_alcoholuses_datacoding_521_f20414_0_0<0)]<-NA4.9 Stress / trauma
UKBvar$V4[which(UKBvar$V2=="Trauma \n stress")]
# CHILDHODD
TT$felt_hated_by_family_member_as_a_childuses_datacoding_532_f20487_0_0[which(TT$felt_hated_by_family_member_as_a_childuses_datacoding_532_f20487_0_0<0)]<-NA
TT$physically_abused_by_family_as_a_childuses_datacoding_532_f20488_0_0[which(TT$physically_abused_by_family_as_a_childuses_datacoding_532_f20488_0_0<0)]<-NA
TT$felt_loved_as_a_childuses_datacoding_532_f20489_0_0[which(TT$felt_loved_as_a_childuses_datacoding_532_f20489_0_0<0)]<-NA
TT$sexually_molested_as_a_childuses_datacoding_532_f20490_0_0[which(TT$sexually_molested_as_a_childuses_datacoding_532_f20490_0_0<0)]<-NA
# TRAUMA
TT$belittlement_by_partner_or_expartner_as_an_adultuses_datacoding_532_f20521_0_0[which(TT$belittlement_by_partner_or_expartner_as_an_adultuses_datacoding_532_f20521_0_0<0)]<-NA
TT$been_in_a_confiding_relationship_as_an_adultuses_datacoding_532_f20522_0_0[which(TT$been_in_a_confiding_relationship_as_an_adultuses_datacoding_532_f20522_0_0<0)]<-NA
TT$physical_violence_by_partner_or_expartner_as_an_adultuses_datacoding_532_f20523_0_0[which(TT$physical_violence_by_partner_or_expartner_as_an_adultuses_datacoding_532_f20523_0_0<0)]<-NA
TT$sexual_interference_by_partner_or_expartner_without_consent_as_an_adultuses_datacoding_532_f20524_0_0[which(TT$sexual_interference_by_partner_or_expartner_without_consent_as_an_adultuses_datacoding_532_f20524_0_0<0)]<-NA
TT$able_to_pay_rentmortgage_as_an_adultuses_datacoding_532_f20525_0_0[which(TT$able_to_pay_rentmortgage_as_an_adultuses_datacoding_532_f20525_0_0<0)]<-NA
TT$been_involved_in_combat_or_exposed_to_warzoneuses_datacoding_533_f20527_0_0[which(TT$been_involved_in_combat_or_exposed_to_warzoneuses_datacoding_533_f20527_0_0<0)]<-NA
TT$victim_of_sexual_assaultuses_datacoding_533_f20531_0_0[which(TT$victim_of_sexual_assaultuses_datacoding_533_f20531_0_0<0)]<-NA
TT$victim_of_sexual_assaultuses_datacoding_533_f20531_0_0[which(TT$victim_of_sexual_assaultuses_datacoding_533_f20531_0_0==2)]<-1
TT$victim_of_physically_violent_crimeuses_datacoding_533_f20529_0_0[which(TT$victim_of_physically_violent_crimeuses_datacoding_533_f20529_0_0<0)]<-NA
TT$victim_of_physically_violent_crimeuses_datacoding_533_f20529_0_0[which(TT$victim_of_physically_violent_crimeuses_datacoding_533_f20529_0_0==2)]<-1
TT$witnessed_sudden_violent_deathuses_datacoding_533_f20530_0_0[which(TT$witnessed_sudden_violent_deathuses_datacoding_533_f20530_0_0<0)]<-NA4.10 Some more info - phase and site
TT$acquisition_protocol_phaseuses_datacoding_778_f25780_2_0[which(is.na(TT$acquisition_protocol_phaseuses_datacoding_778_f25780_2_0))]=5
table(TT$acquisition_protocol_phaseuses_datacoding_778_f25780_2_0)
table(TT$uk_biobank_assessment_centre_f54_2_0, TT$DiscoverySample)
table(TT$uk_biobank_assessment_centre_f54_2_0, TT$ReplicationSample)5 Regress out covariates and calculate fixed effects R2
The phenotypes regressed out for covariates are written in several folders
The main folders for the analysis are: - UKB_phenotypes_reg_Bsln (baseline covariates) and - UKB_phenotypes_reg_Bsln_bdsz (baseline covariate + body size)
Each phenotype is written in an individual tab separated file with 3 columns (no column names) eid eid phenotype. This way we can easily loop on phenotypes in the models by selecting the relevant file.
res=as.data.frame(matrix(NA, ncol=8, nrow = length(UKBvar$V4)) )
colnames(res)=c("Variable", "r0", "rage", "rsex", "rbrain", "rbody","rheight", "rWBMI")
res$Variable=UKBvar$V4
# Regress age sex, brain size
for (var in UKBvar$V4){
m0=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) " ) ), data = TT, na.action = na.exclude)
s0=summary(m0)
m1=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) + Age_MRI " ) ), data = TT, na.action = na.exclude)
s1=summary(m1)
m2=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) + Age_MRI + sexuses_datacoding_9_f31_0_0 " ) ), data = TT, na.action = na.exclude)
s2=summary(m2)
m3=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) + Age_MRI + sexuses_datacoding_9_f31_0_0 + ICV + LThickness + RThickness + LSurfArea + RSurfArea " ) ), data = TT, na.action = na.exclude)
s3=summary(m3)
m4=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) + Age_MRI + sexuses_datacoding_9_f31_0_0 + ICV + LThickness + RThickness + LSurfArea + RSurfArea + standing_height_f50_2_0 " ) ), data = TT, na.action = na.exclude)
s4=summary(m4)
m5=lm(as.formula(paste0( var, " ~ factor(acquisition_protocol_phaseuses_datacoding_778_f25780_2_0) + Age_MRI + sexuses_datacoding_9_f31_0_0 + ICV + LThickness + RThickness + LSurfArea + RSurfArea + standing_height_f50_2_0 + weight_f21002_2_0 + body_mass_index_bmi_f21001_2_0 " ) ), data = TT, na.action = na.exclude)
s5=summary(m5)
# Store residuals of regression
TT[,paste0(var, "_reg")]=scale(residuals(m2))
TT[,paste0(var, "_regB")]=scale(residuals(m3))
TT[,paste0(var, "_regBH")]=scale(residuals(m4))
TT[,paste0(var, "_regBB")]=scale(residuals(m5))
write.table(TT[which(!is.na(TT[,paste0(var)]) ),c("eid", "eid", paste0(var))], paste0("UKB_phenotypes/", var, "_reg", ".txt"), sep="\t", row.names=F, col.names=F , quote=F)
write.table(TT[which(!is.na(TT[,paste0(var, "_reg")]) ),c("eid", "eid", paste0(var, "_reg"))], paste0("UKB_phenotypes_reg/", var, "_reg", ".txt"), sep="\t", row.names=F, col.names=F , quote=F)
write.table(TT[which(!is.na(TT[,paste0(var, "_regB")])),c("eid", "eid", paste0(var, "_regB"))], paste0("UKB_phenotypes_reg_Bsln/", var, "_reg", ".txt"), sep="\t", row.names=F, col.names=F , quote=F)
write.table(TT[which(!is.na(TT[,paste0(var, "_regBH")])),c("eid", "eid", paste0(var, "_regBH"))], paste0("UKB_phenotypes_reg_Bsln_hgt/", var, "_reg", ".txt"), sep="\t", row.names=F, col.names=F , quote=F)
write.table(TT[which(!is.na(TT[,paste0(var, "_regBB")])),c("eid", "eid", paste0(var, "_regBB"))], paste0("UKB_phenotypes_reg_Bsln_bdsz/", var, "_reg", ".txt"), sep="\t", row.names=F, col.names=F , quote=F)
res[which(res$Variable==var),c( "r0", "rage", "rsex", "rbrain", "rbody", "rheight", "rWBMI")]=c(s0$r.squared, s1$r.squared-s0$r.squared, s2$r.squared-s1$r.squared, s3$r.squared-s2$r.squared, s5$r.squared-s3$r.squared, s4$r.squared-s3$r.squared, s5$r.squared-s4$r.squared)
}
# Write fixed effect R2
write.csv(res, "FixedEffects_UKB_15K_Jan2019Update.csv")
#######################
#Write formatted data
write_csv(TT, "UKB_phenotypes_15K_allData_regVars_Jan2019Update.csv")A work by by [Baptiste Couvy-Duchesne] - 23 April 2020