Software: R version 3.6.1
# Prerequisites
loadpkg <- c("xgboost","data.table")
suppressPackageStartupMessages(lapply(loadpkg, function(pkg){
if(!require(pkg,character.only = T)){
install.packages(pkg, dependencies = T)
}
require(pkg,character.only = T)
}))# load models
load_model <- get(load("ald_model.RData"))
# feature information
feature.info <- readRDS("ald_feature_info.rds")
# feature names
predictor.names <- readRDS("ald_predictor_names.rds")The following conventional features are required. Categorical features include sex and smoking status.
feature.info$conventional.feature[,V1]## [1] "sex"
## [2] "age"
## [3] "body mass index"
## [4] "waist-to-hip ratio"
## [5] "alcohol consumption"
## [6] "smoking status"
## [7] "triglycerides"
## [8] "low-density lipoprotein cholesterol"
## [9] "HIGH-density lipoprotein cholesterol"
## [10] "Gamma-glutamyl transferase"The following microbial features are needed. More infomation about GTDB taxonomy can be found at https://gtdb.ecogenomic.org
str(feature.info$microbial.feature) ## List of 10
## $ : chr [1:109] "Absiella sp000163515" "Acetivibrio A ethanolgignens" "Agathobacter sp900317585" "Agathobaculum sp900291975" ...
## $ : chr [1:128] " 14 2 sp001940225" "Agathobacter sp900317585" "An200 sp002160025" "Anaeromassilibacillus sp002159845" ...
## $ : chr [1:276] " 14 2 sp001940225" "Acinetobacter baumannii" "Actinomyces graevenitzii" "Adlercreutzia caecimuris" ...
## $ : chr [1:229] "Absiella innocuum" "Absiella sp000165065" "Acetatifactor muris" "Acinetobacter baumannii" ...
## $ : chr [1:263] " 14 2 sp001940225" "Acetatifactor muris" "Actinomyces graevenitzii" "Acutalibacter sp000435395" ...
## $ : chr [1:262] " 14 2 sp001940225" " 43 108 sp001915545" "Absiella innocuum" "Acetivibrio A ethanolgignens" ...
## $ : chr [1:304] " 43 108 sp001915545" "Absiella dolichum" "Acetatifactor muris" "Acinetobacter baumannii" ...
## $ : chr [1:288] " 14 2 sp001940225" "Acetatifactor muris" "Acinetobacter baumannii" "Actinomyces graevenitzii" ...
## $ : chr [1:332] "Absiella sp000163515" "Absiella sp000165065" "Acetatifactor muris" "Acinetobacter baumannii" ...
## $ : chr [1:248] " 14 2 sp001940225" "Absiella dolichum" "Actinomyces graevenitzii" "Acutalibacter sp000435395" ...To apply the models on new data, the feature names of the corresponding dataset need to be formatted as follows.
lapply(predictor.names, head)## [[1]]
## [1] "Bifidobacterium_sp002742445" "Butyricicoccus_A_porcorum"
## [3] "Prevotella_sp900318795" "Butyricimonas_virosa"
## [5] "Dorea_formicigenerans" "CAG_273_sp000438355"
##
## [[2]]
## [1] "ER4_sp002437735" "Desulfovibrio_piger"
## [3] "Parabacteroides_johnsonii" "CAG_475_sp002449145"
## [5] "Butyricicoccus_A_porcorum" "CAG_83_sp000435975"
##
## [[3]]
## [1] "Desulfovibrio_piger" "CAG_462_sp003489705"
## [3] "Dakarella_massiliensis" "CAG_1138_sp000434675"
## [5] "Prevotella_sp002297965" "Lactobacillus_H_fermentum"
##
## [[4]]
## [1] "CAG_448_sp003150135" "C941_sp900318865"
## [3] "CAG_462_sp003489705" "RC9_sp002393945"
## [5] "Intestinibacter_bartlettii" "Firm_10_sp001603025"
##
## [[5]]
## [1] "CAG_448_sp003150135" "RC9_sp002393945"
## [3] "Desulfovibrio_piger" "UBA11524_sp000437595"
## [5] "CAG_273_sp000437855" "Lactobacillus_H_fermentum"
##
## [[6]]
## [1] "CAG_462_sp003489705" "CAG_448_sp003150135"
## [3] "Desulfovibrio_piger" "Lactobacillus_H_fermentum"
## [5] "Intestinibacter_bartlettii" "TF01_11_sp003524945"
##
## [[7]]
## [1] "UBA11524_sp000437595" "CAG_462_sp003489705"
## [3] "CAG_448_sp003150135" "Prevotella_sp900318795"
## [5] "Desulfovibrio_piger" "Lactobacillus_H_fermentum"
##
## [[8]]
## [1] "CAG_448_sp003150135" "UBA11524_sp000437595"
## [3] "CAG_462_sp003489705" "Eubacterium_E_hallii_A"
## [5] "Intestinibacter_bartlettii" "Faecalicatena_torques"
##
## [[9]]
## [1] "CAG_448_sp003150135" "CAG_462_sp003489705" "Butyricimonas_virosa"
## [4] "Desulfovibrio_piger" "Faecalicatena_torques" "Agathobacter_rectale"
##
## [[10]]
## [1] "RC9_sp002393945" "Dakarella_massiliensis"
## [3] "CAG_273_sp000437855" "CAG_83_sp000435975"
## [5] "C941_sp900318865" "UBA11524_sp000437595"Example of applying a model on new data with required features.
model <- xgb.Booster.complete(load_model[[1]])
# newdata: xgb.DMatrix object with formatted features in predictor.names[[1]]
pred <- predict(model, newdata) # load models
load_model <- get(load("ld_model.RData"))
# feature information
feature.info <- readRDS("ld_feature_info.rds")
# feature names
predictor.names <- readRDS("ld_predictor_names.rds")The following conventional features are required. Categorical features include sex and smoking status.
feature.info$conventional.feature[,V1]## [1] "sex"
## [2] "age"
## [3] "body mass index"
## [4] "waist-to-hip ratio"
## [5] "alcohol consumption"
## [6] "smoking status"
## [7] "triglycerides"
## [8] "low-density lipoprotein cholesterol"
## [9] "HIGH-density lipoprotein cholesterol"
## [10] "Gamma-glutamyl transferase"The following microbial features are needed. More infomation about GTDB taxonomy can be found at https://gtdb.ecogenomic.org
str(feature.info$microbial.feature) ## List of 10
## $ : chr [1:160] "Absiella sp000163515" "Absiella sp000165065" "Acetatifactor sp900066365" "Actinotalea carbonis" ...
## $ : chr [1:112] "Absiella sp000165065" "Actinomyces graevenitzii" "Actinotalea carbonis" "Agathobaculum butyriciproducens" ...
## $ : chr [1:157] "Absiella sp000165065" "Acetatifactor sp003447295" "Acetatifactor sp900066365" "Acetatifactor sp900066565" ...
## $ : chr [1:59] "Absiella sp000163515" "Acetatifactor sp900066365" "Actinomyces graevenitzii" "Agathobaculum butyriciproducens" ...
## $ : chr [1:195] " 51 20 sp001917175" "Absiella innocuum" "Absiella sp000165065" "Acetatifactor sp003447295" ...
## $ : chr [1:143] "Acetatifactor sp900066365" "Actinomyces graevenitzii" "Agathobaculum butyriciproducens" "An200 sp002160025" ...
## $ : chr [1:209] "Absiella innocuum" "Absiella sp000165065" "Acetatifactor sp003447295" "Acetatifactor sp900066365" ...
## $ : chr [1:45] " 43 108 sp001915545" "Acetatifactor sp900066365" "Acetatifactor sp900066565" "Acutalibacter sp000432995" ...
## $ : chr [1:247] "Absiella innocuum" "Absiella sp000163515" "Absiella sp000165065" "Acetatifactor muris" ...
## $ : chr [1:92] "Absiella sp000165065" "Acetatifactor sp900066365" "Actinomyces graevenitzii" "Anaerotignum sp000436415" ...To apply the models on new data, the feature names of the corresponding dataset need to be formatted as follows.
lapply(predictor.names, head)## [[1]]
## [1] "CAG_269_sp003518755" "QALS01_sp003150575"
## [3] "Acetatifactor_sp900066365" "Marseille_P4683_sp900232885"
## [5] "Dakarella_massiliensis" "CAG_1427_sp000435475"
##
## [[2]]
## [1] "CAG_448_sp003150135" "ER4_sp002437735"
## [3] "Streptococcus_sp000187445" "UCG_010_sp003150115"
## [5] "UBA3263_sp001689615" "Ruminococcus_C_sp000437175"
##
## [[3]]
## [1] "UBA5905_sp002431285" "Streptococcus_sp000187445"
## [3] "Sharpea_azabuensis" "CAG_269_sp001915995"
## [5] "Lactobacillus_H_fermentum" "CAG_448_sp003150135"
##
## [[4]]
## [1] "Faecalicatena_torques" "Parabacteroides_johnsonii"
## [3] "Streptococcus_sp000187445" "UBA5905_sp002431285"
## [5] "Agathobaculum_butyriciproducens" "Marseille_P4683_sp900232885"
##
## [[5]]
## [1] "CAG_460_sp000437355" "UBA9475_sp002161235"
## [3] "Acetatifactor_sp900066365" "Marseille_P4683_sp900232885"
## [5] "Anaerotignum_sp000436415" "Dorea_sp900066765"
##
## [[6]]
## [1] "Acetatifactor_sp900066365" "UBA11471_sp000434215"
## [3] "Anaerotignum_sp000436415" "UBA4717_sp002404395"
## [5] "UBA5905_sp002431285" "UBA1375_sp002305795"
##
## [[7]]
## [1] "Streptococcus_sp000187445" "Ruminococcus_A_sp000432335"
## [3] "Lactobacillus_H_fermentum" "Dakarella_massiliensis"
## [5] "Parabacteroides_johnsonii" "CAG_110_sp900315595"
##
## [[8]]
## [1] "Lactobacillus_H_fermentum" "Acetatifactor_sp900066365"
## [3] "Ruminococcus_C_sp000437175" "Ruminococcus_A_sp000432335"
## [5] "Alistipes_sp900021155" "taxa_43_108_sp001915545"
##
## [[9]]
## [1] "CAG_83_sp000435975" "CAG_95_sp000438155"
## [3] "Prevotella_sp900315565" "UC5_1_2E3_sp001304875"
## [5] "Porphyromonas_somerae" "Lactobacillus_H_fermentum"
##
## [[10]]
## [1] "CAG_110_sp000435995" "Acetatifactor_sp900066365"
## [3] "Streptococcus_sp000187445" "Prevotella_maculosa"
## [5] "Ruminococcus_D_sp000434695" "CAG_448_sp003150135"Example of applying a model on new data with required features.
model <- xgb.Booster.complete(load_model[[1]])
# newdata: xgb.DMatrix object with formatted features in predictor.names[[1]]
pred <- predict(model, newdata)