Dear all,
I am submitting this comments about uncertainty on the spirit to improve the draft presented by the AHTEG. It is also in Spanish, because it is my first language, as for many of the 185 experts nominated.
Best regards,
Adriana

Incertidumbre

En la introducción de la hoja de ruta en la sección de "Identificación y consideración de la incertidumbre” se hace referencia a que el protocolo ofrece dos opciones para lidiar con la incertidumbre sobre los niveles de riesgo: 1) solicitar mayor información y/o 2) el dar seguimiento (o monitorear) al OVM en el ambiente receptor. En el documento presentado solo se desarrolla un poco la primera opción, mientras que la segunda, que es en donde puedes generar datos que reduzcan la incertidumbre es olvidada en el resto del documento (monitoreo no visto como manejo de riesgo, sino en este caso, como la obtención de datos que fortalezcan el análisis de riesgo).

La incertidumbre es inherente a la ciencia, no solo al concepto de riesgo. Es comprensible que como humanos haya muchas cosas que no sabemos que desconocemos; incluso he llegado a escuchar a colega mencionando que le preocupa el no saber de que preocuparse, muy legitima preocupación. Sin embargo, en una guía como esta, debiéramos estar orientando a los usuarios en como atender las dos opciones que brinda el Protocolo en su anexo III y no ir más allá de su mandato. No entiendo cuál es la razón del último punto que se presenta en cada uno de los cinco pasos que textualmente dice:

Una consideración de la incertidumbre generada en el paso 1, 2, 3, 4 y 5.. (ver “Identificación y consideración de la incertidumbre”).

¿Cómo puede éste ser un punto independiente a considerar? En cada uno de los otros puntos a considerar es ya tomada en cuenta la incertidumbre que se genera por el muestreo, tamaño de la muestra, tipo de mediciones e instrumentos usados, incertidumbre en el análisis estadístico, etc. Entonces ya es considerada por ejemplo en el primer paso, dentro de cada uno de las consideraciones respecto a la caracterización del OVM, del ambiente receptor y del efecto adverso potencial resultado de la interacción entre el OVM y el ambiente receptor. En los años que llevo realizando análisis de riesgo por la liberación de OVMs al ambiente, nunca he sabido de un método para considerar la incertidumbre por si misma, ésta es considerada dentro de los otros puntos y siempre en relación con la hipótesis de riesgo. En mi opinión este último punto referente a la incertidumbre en cada uno de los cinco pasos debiera ser eliminado; considerando además, como lo menciona el mismo documento, que “actualmente, no existe una definición con acuerdo internacional para la ‘incertidumbre científica’ ni acuerdo internacional sobre reglas generales o directrices para determinar su ocurrencia”.


Uncertainty

Since the introduction of the draft, on the section of “Identification and consideration of uncertainty” is mentioned that the Protocol offers two options to deal with uncertainty: 1) request more information and/or 2) monitoring the living modified organism in the receiving environment.
In the road map only is developed the first option and it seems that the second, where you can obtain data that reduce the uncertainty is forgotten (monitoring not as risk management, but as data obtaining for risk assessment).

Uncertainty is inherent to Science, not only to the concept of risk, I understand that as humans there are many things that we do not know that are unknown, and I have heard a college mentioning that she is worry of not knowing what to be worry about, legitimate concern in my opinion. But in a guide like this, we must orient to the users in how to address the two options that the Protocol on the Anex III mentions, and no more than that. I do not understand the reason of the last point in each of the five steps that textually say:

A consideration of uncertainty arising in step 1, 2, 3 4 and 5 … (see “Identification and consideration of uncertainty” under “Overarching issues in the risk assessment process” above).

How can it be a point to consider by itself? In all the other points to consider there is a part of uncertainty on the method of sampling, size of the sample, measurement instruments, uncertainty of the statistic analysis used, etc… it is already considered for example, for each of the points regarding the characterization of the LMO, regarding the receiving environment, regarding the potential adverse effects resulting from the interaction between the LMO and the receiving environment. In all the years that I have done risk assessment we do not have a method for consider uncertainty analysis by itself, it is inside each of all the other points to consider and always in relation with the risk hypothesis. In my opinion it should be eliminated as an independent point to consider on each of the five steps of the risk assessment. Specially considering, as it is mention on the same draft that to date, “there is no internationally agreed definition of ‘scientific uncertainty’, nor are there internationally agreed general rules or guidelines to determine its occurrence.

Dear all,
General comment:
In some occasions where I have obtained feedback from risk assessors partiuclarly those working in Asia region.
The feedback most commonly comments on the language/terms used in the roadmaps, for example the concept of uncertainty (among others).
Dr. Adriana correctly pointed out, and I quote from the last paragraph that “there is no internationally agreed definition of ‘scientific uncertainty’, nor are there internationally agreed general rules or guidelines to determine its occurrence”. While even in the legal terms, the standard of proof is either on the “balance of probability” or “beyond all reasonable doubts”. It is not an easy task to define for the scientific community on uncertainty. I am not entirely sure whether the AHTEG may want to act as a platform to provide internationally accepted concept of ‘scientific uncertainty”?

Thanks, Dr. Kok Gan Chan, for your precise remarks.
I was perhaps unfortunate in my text, as I did not stress from the very beginning the theoretical possibilities that you listed in your comments. But all of them are contemplated somewhere in the text body. Gene interactions and gene product interactions do happen. The interaction of transgene products, however, is not to be expected as long as they do not belong to the same metabolic pathway. Anyway, if any important interaction (from the strict biosafety point of view) exists in a given stack, it will change either the plant composition or the agronomical behavior. And both aspects are regularly scrutinized by the manufacturer and the regulator. Other unanticipated changes may happen, but it is very hard to predict and presently there in no such a powerful technique to pinpoint these changes before commercial release. Therefore, experiments both in containment and as field trials, are important, and these, too, are regularly performed.
In conclusion, unanticipated interactions may occur, but there are no simple experiments to identify those that slip from theory to practice. And most of them will be, anyway, irrelevant for the safety of the analyzed product. Therefore, I still stick to the suggestions of De Schrijver et al.(2007), as far as a feasible, practical and not too expensive risk assessment is to be done by any regulator.
I very much appreciated participating at this forum and hope that, in spite of been sometimes too incisive, my comments will be somehow useful for improving the AHTEG texts.
Best regards,
Paulo Andrade

Thanks, Dr. Kok Gan Chan, for your precise remarks.
I was perhaps unfortunate in my text, as I did not stress from the very beginning the theoretical possibilities that you listed in your comments. But all of them are contemplated somewhere in the text body. Gene interactions and gene product interactions do happen. The interaction of transgene products, however, is not to be expected as long as they do not belong to the same metabolic pathway. Anyway, if any important interaction (from the strict biosafety point of view) exists in a given stack, it will change either the plant composition or the agronomical behavior. And both aspects are regularly scrutinized by the manufacturer and the regulator. Other unanticipated changes may happen, but it is very hard to predict and presently there in no such a powerful technique to pinpoint these changes before commercial release. Therefore, experiments both in containment and as field trials, are important, and these, too, are regularly performed.
In conclusion, unanticipated interactions may occur, but there are no simple experiments to identify those that slip from theory to practice. And most of them will be, anyway, irrelevant for the safety of the analyzed product. Therefore, I still stick to the suggestions of De Schrijver et al.(2007), as far as a feasible, practical and not too expensive risk assessment is to be done by any regulator.
I very much appreciated participating at this forum and hope that, in spite of been sometimes too incisive, my comments will be somehow useful for improving the AHTEG texts.
Best regards,
Paulo Andrade