calc_excess: Calculation of excess atom fraction

In bramstone/qsip: Quantitative stable isotope probing for microbial ecology

Calculation of excess atom fraction

Description

Calculates fractional isotope incorporation in excess of natural abundances

Usage

calc_excess(
  data,
  tax_id = c(),
  sample_id = c(),
  wads = "wad",
  iso_trt = c(),
  isotope = c(),
  bootstrap = FALSE,
  iters = 999L,
  grouping_cols = c(),
  min_freq = 3,
  correction = TRUE,
  rm_outliers = TRUE,
  non_grower_prop = 0.1,
  nat_abund_13C = 0.01111233,
  nat_abund_15N = 0.003663004,
  nat_abund_18O = 0.002011429
)

Arguments

data	Data as a long-format data.table where each row represents a taxonomic feature within a single fraction. Typically, this is the output from the calc_wad function.
tax_id	Column name specifying unique identifier for each taxonomic feature.
sample_id	Column name specifying unique identifier for each replicate.
wads	Column name specifying weighted average density values. It is possible to change this if preferred but not recommended.
iso_trt	Column name specifying a two-level categorical column indicating whether a sample has been amended with a stable isotope (i.e., is "heavy") or if isotopic composition is at natural abundance (i.e., "light"). Any terms may be applied but care should be taken for these values. If supplied as a factor, calc_excess will take the lowest level as the "light" treatment and the higher level as the "heavy" treatment. Alternatively, if supplied as a character, calc_excess will coerce the column to a factor and with the default behavior wherein the first value in alphabetical order will be assumed to be the lowest factor level (i.e. the "light" treatment).
isotope	Column name specifying the isotope applied to each replicate. For "heavy" samples, these values should be one of "13C", "15N", or "18O".
bootstrap	Whether to generate bootstrapped enrichment values for each taxonomic feature across groups of samples. If TRUE, replicates within specified treatment groupings will be randomly resampled with replacement and resulting WAD values used to generate a distribution of enrichment values for each taxon.
iters	Integer specifying the number of bootstrap iterations to perform.
grouping_cols	Column(s) used to indicate bootstrap resampling groups. Within each group, replicates will be resampled with replacement. Resampling will not occur across groups.
min_freq	Minimum number of replicates a taxonomic feature must occur in to be kept. If treatment grouping columns are specified, frequencies will be assessed at this level. For unlabeled "light" samples, treatment groupings will be ignored when assessing adequate frequency of occurrence.
correction	Whether to apply a correction to fractional enrichment values to ensure a certain proportion are positive.
non_grower_prop	Fractional value applied if correction == TRUE specifying the proportion of the community in each samples assumed to be non-growers and whose median enrichment values will be assumed to be zero. The adjustment necessary to place this median value at zero will be applied as a correction to all enrichment values in the sample.
rm_outlers	Whether or not to remove low fractional enrichment values that are 1.5X lesser than the distance between the 25th quantile and interquartile range. If bootstrap = TRUE, outlier WAD values will be removed prior to resampling and enrichment calculation.

Details

calc_excess automatically averages the isotopically unamended WAD values for each taxonomic feature on the assumption that density values will be identical (or nearly identical) for those samples.

The equations for calculating the molecular weights of taxon i, designated M_{Lab,i} for labeled and M_{Light,i} for unlabeled, are:

M_{Light,i} = 0.496 \cdot G_{i} + 307.691

M_{Lab,i} = \left( \frac{\Delta W}{W_{Light,i}} + 1 \right) \cdot M_{Light,i}

Where

G_{i} = \frac{1}{0.083506} \cdot (W_{Light,i} - 1.646057)

Which indicates the GC content of taxon i based on the density of its DNA when unlabeled

The calculation for the fractional of enrichment of taxon i, A_{i} is:

A_{i} = \frac{M_{Lab,i} - M_{Light,i}}{M_{Heavymax,i} - M_{Light,i}} \cdot (1 - N_{x})

Where

N_{x}: The natural abundance of heavy isotope. Default estimates are: N_{18O} = 0.002000429, N_{13C} = 0.01111233, and N_{15N} = 0.003663004

N_{Heavymax,i}: The highest theoretical molecular weight of taxon i assuming maximum labeling by the heavy isotope

N_{O,Heavymax,i} = (12.07747 + M_{Light,i}) \cdot L

N_{C,Heavymax,i} = (-0.4987282 \cdot G_{i} + 9.974564 + M_{Light,i}) \cdot L

N_{N,Heavymax,i} =( 0.5024851 \cdot G_{i} + 3.517396 + M_{Light,i}) \cdot L

L: The maximum label possible based off the percent of heavy isotope making up the atoms of that element in the labeled treatment

Value

calc_excess returns a data.table where each row represents a taxonomic feature within a single replicate. The following columns are produced: excess atom fraction (eaf). NA values (usually when an organism is present in only the unlabeled or labeled samples) are removed.

Because fraction-level data are being condensed to replicate-level, a list of grouping columns to keep is not necessary unless bootstrapping is specified, wherein those groupings will be used to organize the resampling effort.

Bootstrap functionality produces the columns: 2.5%, 50%, and 97.5% reflecting the median bootstrapped enrichment value and the 95% confidence interval of enrichment values. In addition, probability that a taxon's enrichment was less than zero is represented in the p_val column which is based on the proportion sub-zero bootstrapped observations.

References

Hungate, Bruce, et al. 2015. Quantitative microbial ecology through stable isotope probing. Applied and Environmental Microbiology 81: 7570 - 7581.

Morrissey, Ember, et al. 2018. Taxonomic patterns in nitrogen assimilation of soil prokaryotes. Environmental Microbiology 20: 1112 - 1119.

See Also

calc_wad, wad_wide

Examples

 # Load in example data
 data(example_qsip)

 # relativize sequence abundances (should be done after taxonomic filtering)
 example_qsip[, rel_abund := seq_abund / sum(seq_abund), by = sampleID]

 # calculate weighted average densities
 wads <- calc_wad(example_qsip,
                  tax_id = 'asv_id', sample_id = 'sampleID', frac_id = 'fraction',
                  frac_dens = 'Density.g.ml', frac_abund = 'avg_16S_g_soil',
                  rel_abund = 'rel_abund',
                  grouping_cols = c('treatment', 'isotope', 'iso_trt', 'Phylum'))

 # calculate fractional enrichment in excess of background
 eaf <- calc_excess(wads, tax_id = 'asv_id', sample_id = 'sampleID',
                    iso_trt = 'iso_trt', isotope = 'isotope')

bramstone/qsip documentation built on Nov. 22, 2023, 9:11 p.m.