# Dose Proportionality - NCA

**URL:** <https://discourse.pumas.ai/t/dose-proportionality-nca/987>\
**Category:** Uncategorized\
**Created:** [February 22, 2024, 5:45pm UTC](https://discourse.pumas.ai/t/dose-proportionality-nca/987 "2024-02-22T17:45:05Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![parssh.mehta](https://avatars.discourse-cdn.com/v4/letter/p/8c91f0/32.png) [@parssh.mehta](https://discourse.pumas.ai/u/parssh.mehta)\
**Post date:** [February 22, 2024, 5:45pm UTC](https://discourse.pumas.ai/t/dose-proportionality-nca/987/1 "2024-02-22T17:45:05Z")

</div>

Hi,

I am trying to run Dose Proportionality in Pumas and have the following error. Can you please help with that.

These are the doses in my NCA dataset.

```auto
 Row │ ID COHORT TIME DV route DOSE  
     │ Int64 Int64 Float64 Float64 String3 Int64 
─────┼─────────────────────────────────────────────────
   1 │ 1 1 0.0 0.0 ev 25
   2 │ 7 2 0.0 0.0 ev 50
   3 │ 12 3 0.0 0.0 ev 75
   4 │ 16 4 0.0 0.0 ev 125
   5 │ 20 5 0.0 0.0 ev 250

```

When I run NCA using run\_nca funtion with sigdigits=2, the newly generated _dose_ column (line 16-19) has a dose of 120 instead of the actual dose of 125. Conversely when I use sigdigits=3 it works fine and shows the _dose_ in those line to be 125mg. Not sure why this discrepancy would arise by changing the significant digits which are decimal places?

```auto
Row │ id DOSE dose tlag tmax cmax tlast clast clast_pred auclast kel half_life aucinf_obs aucinf_pred vz_f_obs cl_f_obs vz_f_pred c ⋯ │ String Int64 Int64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 Float64 F ⋯─────┼──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── 1 │ 1 25 25 0.0 1.0 67.0 24.0 6.0 5.9 610.0 0.1 6.9 670.0 670.0 0.37 0.038 0.37 ⋯ 2 │ 2 25 25 0.0 3.0 79.0 24.0 15.0 15.0 1000.0 0.082 8.5 1200.0 1200.0 0.26 0.021 0.26     
   3 │ 3 25 25 0.0 2.0 75.0 24.0 7.4 7.3 760.0 0.1 6.7 830.0 830.0 0.29 0.03 0.29     
   4 │ 4 25 25 0.0 2.0 83.0 24.0 16.0 16.0 920.0 0.068 10.0 1200.0 1200.0 0.32 0.022 0.32     
   5 │ 5 25 25 0.0 2.0 94.0 24.0 26.0 26.0 1300.0 0.057 12.0 1700.0 1700.0 0.25 0.014 0.25 ⋯ 6 │ 6 25 25 0.0 2.0 110.0 24.0 18.0 18.0 1200.0 0.075 9.3 1500.0 1400.0 0.23 0.017 0.23     
   7 │ 7 50 50 0.0 4.0 190.0 24.0 37.0 36.0 2100.0 0.073 9.6 2600.0 2600.0 0.27 0.019 0.27     
   8 │ 8 50 50 0.0 1.0 160.0 24.0 17.0 17.0 1500.0 0.087 7.9 1600.0 1600.0 0.35 0.03 0.35     
   9 │ 9 50 50 0.0 1.0 190.0 24.0 35.0 35.0 2200.0 0.07 9.9 2700.0 2700.0 0.27 0.019 0.27 ⋯ 10 │ 10 50 50 0.0 3.0 170.0 24.0 24.0 24.0 1900.0 0.091 7.6 2200.0 2200.0 0.25 0.023 0.25     
  11 │ 11 50 50 0.0 2.0 160.0 24.0 19.0 19.0 1600.0 0.091 7.6 1800.0 1800.0 0.3 0.027 0.3      
  ⋮ │ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋱ 15 │ 15 75 75 0.0 4.0 290.0 24.0 89.0 88.0 4100.0 0.064 11.0 5500.0 5500.0 0.21 0.014 0.21     
  16 │ 16 125 120 0.0 0.75 390.0 24.0 47.0 46.0 4100.0 0.094 7.4 4600.0 4600.0 0.29 0.027 0.29 ⋯ 17 │ 17 125 120 0.0 1.0 520.0 24.0 80.0 79.0 5500.0 0.076 9.2 6600.0 6600.0 0.25 0.019 0.25     
  18 │ 18 125 120 0.0 1.0 520.0 24.0 130.0 130.0 6300.0 0.051 13.0 8800.0 8800.0 0.28 0.014 0.28     
  19 │ 19 125 120 0.0 2.0 540.0 24.0 140.0 140.0 7000.0 0.055 13.0 9500.0 9500.0 0.24 0.013 0.24     
  20 │ 20 250 250 0.0 3.0 680.0 24.0 97.0 97.0 7700.0 0.088 7.9 8900.0 8800.0 0.32 0.028 0.32 ⋯ 21 │ 21 250 250 0.0 2.0 820.0 24.0 170.0 170.0 9300.0 0.063 11.0 12000.0 12000.0 0.33 0.021 0.33     
  22 │ 22 250 250 0.0 3.0 910.0 24.0 230.0 240.0 12000.0 0.061 11.0 16000.0 16000.0 0.26 0.016 0.26     
  23 │ 23 250 250 0.0 2.0 750.0 24.0 130.0 130.0 8400.0 0.07 9.9 10000.0 10000.0 0.35 0.024 0.35     
  24 │ 24 250 250 0.0 0.75 1200.0 24.0 210.0 210.0 14000.0 0.068 10.0 17000.0 17000.0 0.22 0.015 0.22 ⋯ 25 │ 25 250 250 0.0 3.0 920.0 24.0 150.0 150.0 10000.0 0.073 9.5 12000.0 12000.0 0.28 0.02 0.28     

```

This error then causes differences in the calculation of dose proportionality.

```auto
Dose linearity pairwise ratio test
Variable: cmax
────────────────────────────────────────────────────
            Ratio Estimate low CI 90% high CI 90%
────────────────────────────────────────────────────
50 vs. 25 2.0 2.09008 1.7942 2.43475
75 vs. 25 3.0 3.48504 2.93812 4.13375
125 vs. 25 5.0 5.8609 4.83306 7.10732
250 vs. 25 10.0 10.3449 8.6096 12.4299 
────────────────────────────────────────────────────

vs (wrong interpretation)

Dose linearity pairwise ratio test
Variable: cmax
────────────────────────────────────────────────────
            Ratio Estimate low CI 90% high CI 90%
────────────────────────────────────────────────────
50 vs. 25 2.0 2.07577 1.77179 2.4319
75 vs. 25 3.0 3.45615 2.87823 4.1501
120 vs. 25 4.8 5.84529 4.78665 7.13806
250 vs. 25 10.0 10.3555 8.51974 12.5869
────────────────────────────────────────────────────

```

---

<div class="post-metadata">

**Author:** ![andreasnoack](https://yyz2.discourse-cdn.com/flex030/user_avatar/discourse.pumas.ai/andreasnoack/32/22_2.png) [@andreasnoack](https://discourse.pumas.ai/u/andreasnoack)\
**Post date:** [February 27, 2024, 7:33am UTC](https://discourse.pumas.ai/t/dose-proportionality-nca/987/2 "2024-02-27T07:33:15Z")

</div>

This is [how rounding to significant digits](https://en.wikipedia.org/wiki/Significant_figures) work

```julia
julia> round(125, sigdigits=2)
120.0

```

This is not specific to Julia. Here is R

```r
> signif(125, 2)
[1] 120

```

If you just want to round digits after the decimal point then you should use the `digits` keyword

```auto
julia> round(125, digits=2)
125.0

```