Taxa filtering

• MOMS PI 논문 조건 사항 (method - Filtering out low-abundant taxa)
  • 5% of the profiles exhibited an abundance of at least 1%, and
  • at least 15% of profiles exhibited an abundance of at least 0.1%.

조건1 - 5% of the profiles exhibited an abundance of at least 1%

taxa2 <- prop.table(as.matrix(taxa), margin = 1)
ttaxa <- as.data.frame(t(taxa2))

for (i in 1:1851) {
  ttaxa[,i] <- ifelse(ttaxa[,i] >= 0.01, 1, 0)
}

ttaxa$rowMean <- rowMeans(ttaxa)
summary(ttaxa$rowMean)

ttaxa <- ttaxa[ttaxa$rowMean >= 0.05, ]

row.names(ttaxa) -> a

taxa3 <- subset(taxa, select = a)

• dataframes

  • taxa = momspi vaginal taxa table (frequency table)
    • col = taxa (1108 var)
    • row = sample (1851 obs)
  • taxa2 = momspi vaginal taxa table (abundance table)
  • ttaxa = taxa transpose

• 방법

  1. abundance 셀이 0.01 이상인 셀을 1로, 이하인 셀을 0으로 값 변환
  2. taxa 별로 평균값 계산
  3. 평균값이 0.05 이상인 taxa만 남기기
  4. 해당 taxa의 frequnecy 값 가져오기

조건2 - at least 15% of profiles exhibited an abundance of at least 0.1%

ttaxa <- as.data.frame(t(taxa3))
for (i in 1:1851) {
  ttaxa[,i] <- ifelse(ttaxa[,i] >= 0.001, 1, 0)
}

ttaxa$rowMean <- rowMeans(ttaxa)
summary(ttaxa$rowMean)

ttaxa <- ttaxa[ttaxa$rowMean >= 0.15, ]

row.names(ttaxa) -> a

taxa3 <- subset(taxa, select = a)

• 방법
  • 조건1의 방법과 동일하게 진행

Log transformation

Cytokine log tranformation

for (i in 1:27){
  cyto[,i] <- log(cyto[,i])
}

• 각 column 별 값 log로 변환

Taxa table log transformation