hd_normalize() normalizes the data by scaling them and removing their batch effects.
It removes the batch effects and scales or centers the data. To remove batch effects, it uses the
remove_batch_effects(), that utilizes limma package. For scaling, it uses the scale()
from base R.
Usage
hd_normalize(
dat,
metadata = NULL,
center = TRUE,
scale = TRUE,
batch = NULL,
batch2 = NULL
)Arguments
- dat
An HDAnalyzeR object or a dataset in wide format and sample ID as its first column.
- metadata
A dataset containing the metadata information. If a HDAnalyzeR object is provided, this parameter is not needed.
- center
A logical value indicating whether to center the data. Default is TRUE.
- scale
A logical value indicating whether to scale the data. Default is TRUE.
- batch
The metadata column containing the batch information. In order to correct for batch effects, this parameter should be provided. Default is NULL.
- batch2
The metadata column containing the second batch information if available. Default is NULL.
Details
You can read more about the scaling and centering methods in the documentation of the scale() function in the base package,
as well as about the method for removing batch effects in the documentation of the removeBatchEffect() function in the limma package.
Examples
# Create the HDAnalyzeR object providing the data and metadata
hd_object <- hd_initialize(example_data, example_metadata)
# Center data
scaled_dat <- hd_normalize(hd_object, center = TRUE, scale = FALSE)
scaled_dat$data
#> # A tibble: 586 × 101
#> DAid AARSD1 ABL1 ACAA1 ACAN ACE2 ACOX1 ACP5 ACP6 ACTA2
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 DA00001 0.259 0.949 0.697 -0.532 0.827 -1.42 0.612 1.02 1.20
#> 2 DA00002 -1.71 -0.563 -1.83 -1.02 -0.102 -1.40 -0.278 0.168 -0.813
#> 3 DA00003 NA NA NA 0.424 NA -0.167 0.447 NA NA
#> 4 DA00004 0.278 1.57 0.683 NA 0.593 NA -0.0839 -0.551 0.0892
#> 5 DA00005 1.88 3.24 -0.882 -0.165 -1.86 -1.08 -0.898 0.0236 1.12
#> 6 DA00006 3.70 -0.628 -2.75 -0.721 0.600 -1.22 -0.305 -0.606 -0.840
#> 7 DA00007 NA NA 2.95 0.117 2.21 2.12 0.548 1.12 0.398
#> 8 DA00008 -0.351 -0.998 -1.56 0.416 -1.03 -0.800 -0.549 -1.96 -0.0901
#> 9 DA00009 1.26 1.53 -1.46 -1.43 -0.533 1.21 0.562 -1.16 -1.41
#> 10 DA00010 -1.30 -0.596 -1.92 -1.60 -1.02 -0.801 0.765 -1.04 0.427
#> # ℹ 576 more rows
#> # ℹ 91 more variables: ACTN4 <dbl>, ACY1 <dbl>, ADA <dbl>, ADA2 <dbl>,
#> # ADAM15 <dbl>, ADAM23 <dbl>, ADAM8 <dbl>, ADAMTS13 <dbl>, ADAMTS15 <dbl>,
#> # ADAMTS16 <dbl>, ADAMTS8 <dbl>, ADCYAP1R1 <dbl>, ADGRE2 <dbl>, ADGRE5 <dbl>,
#> # ADGRG1 <dbl>, ADGRG2 <dbl>, ADH4 <dbl>, ADM <dbl>, AGER <dbl>, AGR2 <dbl>,
#> # AGR3 <dbl>, AGRN <dbl>, AGRP <dbl>, AGXT <dbl>, AHCY <dbl>, AHSP <dbl>,
#> # AIF1 <dbl>, AIFM1 <dbl>, AK1 <dbl>, AKR1B1 <dbl>, AKR1C4 <dbl>, …
round(mean(scaled_dat$data$ABL1, na.rm = TRUE), 4) # should be 0
#> [1] 0
round(sd(scaled_dat$data$ABL1, na.rm = TRUE), 4)
#> [1] 1.3846
# Center and scale data (z-score scaling)
scaled_dat <- hd_normalize(hd_object, center = TRUE, scale = TRUE)
scaled_dat$data
#> # A tibble: 586 × 101
#> DAid AARSD1 ABL1 ACAA1 ACAN ACE2 ACOX1 ACP5 ACP6 ACTA2
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 DA00001 0.240 0.685 0.498 -0.753 0.722 -1.39 0.800 0.991 1.16
#> 2 DA00002 -1.58 -0.406 -1.30 -1.45 -0.0885 -1.37 -0.364 0.163 -0.786
#> 3 DA00003 NA NA NA 0.600 NA -0.163 0.584 NA NA
#> 4 DA00004 0.257 1.14 0.488 NA 0.517 NA -0.110 -0.536 0.0862
#> 5 DA00005 1.74 2.34 -0.629 -0.233 -1.62 -1.06 -1.18 0.0230 1.08
#> 6 DA00006 3.42 -0.453 -1.96 -1.02 0.523 -1.19 -0.399 -0.590 -0.812
#> 7 DA00007 NA NA 2.11 0.165 1.93 2.08 0.717 1.09 0.385
#> 8 DA00008 -0.325 -0.721 -1.12 0.589 -0.898 -0.783 -0.719 -1.90 -0.0871
#> 9 DA00009 1.17 1.11 -1.04 -2.03 -0.464 1.18 0.735 -1.13 -1.36
#> 10 DA00010 -1.20 -0.431 -1.37 -2.27 -0.889 -0.784 1.00 -1.01 0.413
#> # ℹ 576 more rows
#> # ℹ 91 more variables: ACTN4 <dbl>, ACY1 <dbl>, ADA <dbl>, ADA2 <dbl>,
#> # ADAM15 <dbl>, ADAM23 <dbl>, ADAM8 <dbl>, ADAMTS13 <dbl>, ADAMTS15 <dbl>,
#> # ADAMTS16 <dbl>, ADAMTS8 <dbl>, ADCYAP1R1 <dbl>, ADGRE2 <dbl>, ADGRE5 <dbl>,
#> # ADGRG1 <dbl>, ADGRG2 <dbl>, ADH4 <dbl>, ADM <dbl>, AGER <dbl>, AGR2 <dbl>,
#> # AGR3 <dbl>, AGRN <dbl>, AGRP <dbl>, AGXT <dbl>, AHCY <dbl>, AHSP <dbl>,
#> # AIF1 <dbl>, AIFM1 <dbl>, AK1 <dbl>, AKR1B1 <dbl>, AKR1C4 <dbl>, …
round(mean(scaled_dat$data$ABL1, na.rm = TRUE), 4) # should be 0
#> [1] 0
round(sd(scaled_dat$data$ABL1, na.rm = TRUE), 4) # should be 1
#> [1] 1
# Center, scale and remove batch effects
scaled_dat <- hd_normalize(hd_object, batch = "Cohort")
scaled_dat$data
#> # A tibble: 586 × 101
#> DAid AARSD1 ABL1 ACAA1 ACAN ACE2 ACOX1 ACP5 ACP6 ACTA2 ACTN4
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 DA000… 0.104 0.476 0.391 -0.779 0.544 -1.46 0.701 0.915 0.985 0.516
#> 2 DA000… -1.74 -0.656 -1.43 -1.48 -0.286 -1.44 -0.471 0.0838 -1.02 -0.761
#> 3 DA000… NA NA NA 0.574 NA -0.232 0.484 NA NA NA
#> 4 DA000… 0.121 0.944 0.380 NA 0.335 NA -0.215 -0.618 -0.122 -0.486
#> 5 DA000… 1.62 2.20 -0.746 -0.259 -1.86 -1.13 -1.29 -0.0569 0.903 -0.103
#> 6 DA000… 3.32 -0.705 -2.09 -1.05 0.341 -1.26 -0.507 -0.672 -1.05 NA
#> 7 DA000… NA NA 2.02 0.140 1.78 2.01 0.618 1.01 0.186 -0.388
#> 8 DA000… -0.468 -0.983 -1.24 0.563 -1.12 -0.854 -0.828 -1.99 -0.300 -1.60
#> 9 DA000… 1.04 0.915 -1.16 -2.05 -0.671 1.12 0.636 -1.21 -1.61 -1.50
#> 10 DA000… -1.36 -0.681 -1.49 -2.30 -1.11 -0.855 0.903 -1.10 0.214 0.136
#> # ℹ 576 more rows
#> # ℹ 90 more variables: ACY1 <dbl>, ADA <dbl>, ADA2 <dbl>, ADAM15 <dbl>,
#> # ADAM23 <dbl>, ADAM8 <dbl>, ADAMTS13 <dbl>, ADAMTS15 <dbl>, ADAMTS16 <dbl>,
#> # ADAMTS8 <dbl>, ADCYAP1R1 <dbl>, ADGRE2 <dbl>, ADGRE5 <dbl>, ADGRG1 <dbl>,
#> # ADGRG2 <dbl>, ADH4 <dbl>, ADM <dbl>, AGER <dbl>, AGR2 <dbl>, AGR3 <dbl>,
#> # AGRN <dbl>, AGRP <dbl>, AGXT <dbl>, AHCY <dbl>, AHSP <dbl>, AIF1 <dbl>,
#> # AIFM1 <dbl>, AK1 <dbl>, AKR1B1 <dbl>, AKR1C4 <dbl>, AKT1S1 <dbl>, …