Package 'demic'

Compares contig subset x against contig subset y

Description

Compares contig subset x against contig subset y

Usage

compare_contig_subsets(
  est_ptrs_x,
  est_ptrs_y,
  pipeline_x,
  pipeline_y,
  cor_cutoff,
  max_cor
)

Arguments

est_ptrs_x	PTR estimates from contig subset x
est_ptrs_y	PTR estimates from contig subset y
pipeline_x	pipeline for contig subset x
pipeline_y	pipeline for contig subset y
cor_cutoff	the correlation cutoff
max_cor	the max correlation

Value

a named list including the est_ptr dataframe and a max_cor value

• sample: sample

• est_ptr: PTR estimate

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

max_cor: the max correlation achieved

---

Compares sample subset x against sample subset y

Description

Compares sample subset x against sample subset y

Usage

compare_sample_subsets(
  est_ptrs_x,
  est_ptrs_y,
  pipeline_x,
  pipeline_y,
  cor_cutoff,
  max_cor
)

Arguments

est_ptrs_x	PTR estimates from sample subset x
est_ptrs_y	PTR estimates from sample subset y
pipeline_x	pipeline for sample subset x
pipeline_y	pipeline for sample subset y
cor_cutoff	the correlation cutoff
max_cor	the max correlation

Value

a named list including the est_ptr dataframe and a max_cor value

• sample: sample

• est_ptr: PTR estimate

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

---

A function for data frame integration

Description

A function for data frame integration

Usage

consist_transfer(x, y, i)

Arguments

x	first data frame
y	second data frame
i	'sample' column

Value

a data frame with the other column as mean or max of that in the original two

---

A function to return the first dimension of PCA on an input matrix

Description

A function to return the first dimension of PCA on an input matrix

Usage

contig_pca(X)

Arguments

X	a matrix to undergo PCA

Value

first dimension of the PCA results

---

Contig Cluster 1

Description

Data associated with DEMIC paper (on SourceForge)

Usage

ContigCluster1

Format

ContigCluster1

A data frame with 120,897 rows and 5 columns:

log_cov

Log Coverage for Sliding Windows over Contigs

GC_content

GC Content for Sliding Windows over Contigs

sample

Sample Name

contig

Contig Name

length

Length of Contig

Source

https://sourceforge.net/projects/demic/files/

---

Contig Cluster 2

Description

Data associated with DEMIC paper (on SourceForge)

Usage

ContigCluster2

Format

ContigCluster2

A data frame with 66,735 rows and 5 columns:

log_cov

Log Coverage for Sliding Windows over Contigs

GC_content

GC Content for Sliding Windows over Contigs

sample

Sample Name

contig

Contig Name

length

Length of Contig

Source

https://sourceforge.net/projects/demic/files/

---

Determine the majority orientation of the input PTR estimates correlations

Description

Determine the majority orientation of the input PTR estimates correlations

Usage

cor_diff(Z)

Arguments

Z	a vector of values

Value

a minor subset, where each value has the same orientation

---

A function for data frame transfer

Description

A function for data frame transfer

Usage

df_transfer(x, y)

Arguments

x	first data frame with six columns
y	second data frame with six columns

Value

a data frame with the same six columns but integrated info

---

Estimate PTRs using all input data as well as using subsets of contigs and samples

Description

Estimate PTRs using all input data as well as using subsets of contigs and samples

Usage

est_ptr(X)

Arguments

X	dataframe with coverage matrix (column names: "log_cov", "GC_content", "sample", "contig", "length")

Value

named list with results from all three methods all_ptr dataframe with the estimated PTRs on success, null otherwise

• est_ptr: estimated PTR values

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

contigs_ptr dataframe with the estimated PTRs on success, null otherwise

• est_ptr: estimated PTR values

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

samples_ptr dataframe with the estimated PTRs on success, null otherwise

• est_ptr: estimated PTR values

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

Examples

est_ptrs_001 <- est_ptr(max_bin_003)
est_ptrs_001

---

Tries up to max_attempts times to compare each permutation of removing random subsets of contigs/samples from X, and returns the PTR estimate if a valid one comes back from the comparisons

Description

Requires a minimum of 2 * num_subsets contigs/samples

Usage

est_ptr_on(X, subset_on, max_attempts = 10, num_subsets = 3, cor_cutoff = 0.98)

Arguments

X	cov3 dataframe
subset_on	either "contig" or "sample"
max_attempts	max number of attempts to find a valid ptr estimate
num_subsets	number of subsets to split contigs/samples into
cor_cutoff	minimum correlation coefficient to accept PTR estimate

Value

est_ptrs dataframe on success, null otherwise

• est_ptr: estimated PTR values

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

Examples

est_ptrs_001_on_contigs <- est_ptr_on(max_bin_003, "contig", num_subsets = 5)
est_ptrs_001_on_contigs

est_ptrs_001_on_samples <- est_ptr_on(max_bin_003, "sample")
is.null(est_ptrs_001_on_samples)

---

Estimates PTRs based on the whole input dataset

Description

Estimates PTRs based on the whole input dataset

Usage

est_ptr_on_all(X)

Arguments

X	cov3 dataframe

Value

est_ptrs dataframe on success, null otherwise

• est_ptr: estimated PTR values

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

Examples

est_ptrs_001 <- est_ptr_on_all(max_bin_003)
est_ptrs_001

---

Get PTR estimates for output of the core pipeline on a subset of data

Description

Get PTR estimates for output of the core pipeline on a subset of data

Usage

est_ptrs_subset(p)

Arguments

p	is the pipeline named list

Value

a dataframe

• sample: sample

• est_ptr: PTR estimate

• coefficient: coefficient of linear regression

• pValue: p-value of linear regression

• cor: correlation coefficient

• correctY: corrected coverage

---

A function for sample filtration Input requirements: 1. have values in more than half of the contigs 2. average log2(cov) > 0 in all these contigs

Description

A function for sample filtration Input requirements: 1. have values in more than half of the contigs 2. average log2(cov) > 0 in all these contigs

Usage

filter_sample(Z, avg_cutoff, cutoff_ratio)

Arguments

Z	a matrix
avg_cutoff	threshold of average
cutoff_ratio	threshold of ratio

Value

the coefficient and p value of linear regression

---

Generate a variety of stats on PTR estimates for a given dataset

Description

Generate a variety of stats on PTR estimates for a given dataset

Usage

get_eptr_stats(X, iterations = 30)

Arguments

X	cov3 dataframe
iterations	number of iterations to run

Value

named list of stats on PTR estimates

• all_sd: standard deviation of PTR estimates from all method

• all_mean: mean of PTR estimates from all method

• contigs_sd: standard deviation of PTR estimates from contigs method

• contigs_mean: mean of PTR estimates from contigs method

• samples_sd: standard deviation of PTR estimates from samples method

• samples_mean: mean of PTR estimates from samples method

Examples

stats <- get_eptr_stats(max_bin_001[max_bin_001$sample %in% c('Akk0_001', 'Akk1_001'), ], 2)
stats

---

A function for iteration of pipeline until convergence

Description

A function for iteration of pipeline until convergence

Usage

iterate_pipelines(Z)

Arguments

Z	a matrix of coverages

Value

a named list

• samples: vector of final filtered samples

• correct_ys: matrix of sample, contig and corrected coverages

• pc1: matrix of contig and PC1 values

• pc1_range: vector of PC1 range

• samples_y: samples filtered for reliable coverage

---

A convenient function for KS test of uniform distribution

Description

A convenient function for KS test of uniform distribution

Usage

ks(x)

Arguments

x	a vector without NA

Value

the p value of KS test

---

A convenient function for ordinary linear regression on two vectors

Description

A convenient function for ordinary linear regression on two vectors

Usage

lm_column(x, y)

Arguments

x	first vector
y	second vector

Value

the coefficient and p value of linear regression

---

Run mixed linear model with random effect using lme4

Description

Run mixed linear model with random effect using lme4

Usage

lme4_model(X)

Arguments

X	input data frame

Value

a dataframe

---

MaxBin2 Cluster 001

Description

Generated by PyCov3 on simulated test data

Usage

max_bin_001

Format

max_bin_001

A data frame with 79,740 rows and 5 columns:

log_cov

Log Coverage for Sliding Windows over Contigs

GC_content

GC Content for Sliding Windows over Contigs

sample

Sample Name

contig

Contig Name

length

Length of Contig

Source

https://sourceforge.net/projects/demic/files/

---

MaxBin2 Cluster 002

Description

Generated by PyCov3 on simulated test data

Usage

max_bin_002

Format

max_bin_002

A data frame with 148,638 rows and 5 columns:

log_cov

Log Coverage for Sliding Windows over Contigs

GC_content

GC Content for Sliding Windows over Contigs

sample

Sample Name

contig

Contig Name

length

Length of Contig

Source

https://sourceforge.net/projects/demic/files/

---

MaxBin2 Cluster 003

Description

Generated by PyCov3 on simulated test data

Usage

max_bin_003

Format

max_bin_003

A data frame with 124,578 rows and 5 columns:

log_cov

Log Coverage for Sliding Windows over Contigs

GC_content

GC Content for Sliding Windows over Contigs

sample

Sample Name

contig

Contig Name

length

Length of Contig

Source

https://sourceforge.net/projects/demic/files/

---

A function representing the pipeline of four steps including GC bias correction, sample filtration, PCA and contig filtration

Description

A function representing the pipeline of four steps including GC bias correction, sample filtration, PCA and contig filtration

Usage

pipeline(Y, i)

Arguments

Y	a matrix of coverages
i	cutoff of filtering samples changes according to parameter i; i=1, cutoffRatio is 0.5; i=2, cutoffRatio is 1 as contig is clean

Value

a named list

• samples: final list of filtered samples

• correct_ys: dataframe with correct Y values per contig/sample

• pc1: PC1 results of PCA per contig

• pc1_range: range of PC1

• samples_y: samples filtered for reliable coverage

---

A function for reshape to facilitate PCA, removing all contigs with missing values for designated samples

Description

A function for reshape to facilitate PCA, removing all contigs with missing values for designated samples

Usage

reshape_filtered(samples_filtered, Z)

Arguments

samples_filtered	a vector of samples
Z	a matrix of coverage

Value

a reshaped matrix of coverage

---

A function to remove outlier contigs using KS test

Description

A function to remove outlier contigs using KS test

Usage

select_by_ks_test(sort_values)

Arguments

sort_values

a vector of sorted values

Value

a vector with all values following a uniform distribution

---

A function to test whether the result is reasonable

Description

A function to test whether the result is reasonable

Usage

test_reasonable(a, b)

Arguments

a	first vector of values
b	second vector of values

Value

the test result

---

Verify that the input dataframe/matrix is valid

Description

Verify that the input dataframe/matrix is valid

Usage

verify_input(X)

Arguments

X	dataframe/matrix with cov3 information

---