[Vp-integration-subgroup] Unit standardization section

[Dobrovolny, Hana]

Hi Jacob,


Here's what I suggest as a response to the reviewer:


Unit standardization: The conversion from PFU or TCID50 to individual virions is likely to differ across viruses – are the authors focused on COVID here? Are the authors advocating for a standard conversion factor? It is not clear what the purpose of this discussion is. As the authors mention, different scales require different units. Even at a single scale, different models may require different units for numerical reasons. It is not clear what the authors are advocating for here.


We agree that the conversion from pfu or TCID50 to individual virions is likely to differ across viruses. The conversion estimates cited in the manuscript are for influenza, and are, to our knowledge, the only attempts to estimate the conversion factor for any virus. We have clarified that the conversion estimates are for influenza.


We are not advocating for a standard conversion factor, since the conversion factor likely depends not only on the specific virus, but also conditions such as temperature and pH, which are known to affect viral infectivity (Rowell and Dobrovolny, 2020, Heumann et al, 2021). Rather, we are advocating for development of new viral measurement techniques, that can more reliably quantify the number of infectious virus present in a sample. We have added this discussion to the text of the manuscript.


More generally, we are advocating for clarity in the use of units. Sometimes models are used (and published) without specifying the units used for simulation parameters --- this is a practice that needs to be corrected. Tools such as ClinicalUnitMapping.Com can also be helpful in assuring that standardized units are used for simulation parameters and are converted appropriately when different scales or units are needed. This discussion has also been added to the manuscript.


[1] Rowell, CER, and Dobrovolny, HM (2020), Energy Requirements for Loss of Viral Infectivity, Food and Environmental Virology, 12, 281-294, doi:10.1007/s12560-020-09439-9


[2] Heumann, R, Duncan, C, Stobart, CC, Kaschner, S (2021) Dynamical Differences in Respiratory Syncytial Virus, Bulletin of Mathematical Biology, 84(1), 11 doi: 10.1007/s11538-021-00971-8


And changes to the manuscript:


Unfortunately, units of measure are not yet standardized, an open problem despite many
attempts to resolve it by multiple standardization bodies such as IEEE, CDCIC, and NIST. One indication of the severity of the problem is that a Github search for "unit conversion" shows over a thousand results. Another good example of the severity of the problem is ClincialTrials.Gov that aggregates quantitative data from around the world and this database shows over 24K different units of measure (Barhak and Schertz, 2019). One attempt at solving this standardization issue using machine learning is ClinicalUnitMapping.com, yet this project requires more effort.

Unit mismatches become particularly problematic when trying to integrate models across
different spatial or time scales. For example, intracellular processes occur on micrometer spatial scales and seconds to minutes time scales. An in-host, tissue-level model of infection processes operates at millimeter to centimeter distances and hour to day time scales. When trying to integrate the two into a single multiscale model, care must be taken to ensure appropriate conversion of units when transferring output of one model as input to the other. This broad range of spatial and temporal scales can also cause computational problems requiring development of new algorithms to make computation more efficient across multiple scales (Jung and Sugita, 2017). While it is impossible to avoid having to convert units, we are advocating for clarity in the use of units. Sometimes models are used (and published) without specifying the units used for simulation parameters --- this is a practice that needs to be corrected. Tools such as ClinicalUnitMapping.Com can also be helpful in assuring that standardized units are used for simulation parameters and are converted appropriately when different scales or units are needed. This discussion has also been added to the manuscript.

Lack of standard units for measurement of infectious virions is particularly problematic when trying to develop stochastic viral models. Stochastic models often require that we track individual infectious viral particles, yet it is not clear how the typical viral titer units of TCID 50 /ml and pfu/ml convert to individual virions. Two attempts have been made to estimate the conversion factor, both for influenza, resulting in estimates of 1 TCID 50 /mL of nasal wash corresponding to 10 2 -10 5(Handel et. al., 2007) or 3x10 4 -3x10 5 (Perelson et. al., 2012) virions at the site of infection. Such order of magnitude uncertainty in unit conversion makes it difficult to develop accurate model representations of viral infections. We are not advocating for a standard conversion factor, since the conversion factor likely depends not only on the specific virus, but also conditions such as temperature and pH, which are known to affect viral infectivity (Rowell and Dobrovolny, 2020, Heumann et al, 2021). Rather, we are advocating for development of new viral measurement techniques, that can more reliably quantify the number of infectious virus present in a sample.



Feel free to edit, or add to as you think might be needed (Jacob or anyone else who has thoughts on this).


Hana

*******************************************************
Dr. Hana Dobrovolny
Associate Professor of Biophysics
Texas Christian University
TCU Box 298840
Fort Worth, TX 76129

phone: (817) 257-6379 fax: (817) 257-7742
email: h.dobrovolny at tcu.edu
*******************************************************


________________________________
From: Jacob Barhak <jacob.barhak at gmail.com>
Sent: December 20, 2021 5:24 AM
To: Dobrovolny, Hana
Cc: vp-integration-subgroup at lists.simtk.org; vp-reproduce-subgroup at lists.simtk.org
Subject: Unit standardization section


[EXTERNAL EMAIL WARNING] DO NOT CLICK LINKS or open attachments unless you recognize the sender and know the content is safe.

Hi Hana,

You and I are the main people contributing to the unit standardization section in the paper, therefore we should try to answer reviewer 2.
Here is the review:

"
Unit standardization: The conversion from PFU or TCID50 to individual virions is likely to differ across viruses – are the authors focused on COVID here? Are the authors advocating for a standard conversion factor? It is not clear what the purpose of this discussion is. As the authors mention, different scales require different units. Even at a single scale, different models may require different units for numerical reasons. It is not clear what the authors are advocating for here.
"

I think reviewer is asking for clarity what we are suggesting as a solution - I believe that units should be well standardized so a machine will be able to comprehend those - this is why I started ClinicalUnitMapping.Com

I think the reviewer is asking for a proposed solution the problem you write about - will a tool such as ClinicalUnitMapping.Com help resolve your problem if it reaches maturity or do you have other solutions in mind we can propose to answer the reviewer?

If you wish to discuss more details, I will be happy to virtually meet

             Jacob






-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.simtk.org/pipermail/vp-integration-subgroup/attachments/20211230/290d4091/attachment.html>