68 
Slide 4 
ESTIMATING PROBABILITIES FOR ESCAPE, SURVIVAL, AND MANIFESTATION OF 
ADVERSE CONSEQUENCES OF RECOMBINANT DNA MOLECULES 
Probability of escape from the laboratory depends on: 
1. Personnel --tra i ning and use of acceptable procedures 
2. Faci 1 i t ies--Pl vs P2 vs P3 vs P4 
3. Existence and implementation of accident plan 
4. Ninber of investigators using the techniques 
Probability of survival and perpetuation depends on: 
1. Nature of bacterial host and cloning vector 
2. Ecological niches occupiable by original and chimeric hosts 
3. Transmi ss i bi 1 i ty of recombinant DNA to other bacteria 
4. Selective advantages or disadvantages conferred by foreign DNA 
Probability of manifestation of adverse consequences depends on: 
1. Source of foreign DNA 
2. Expression of foreign DNA 
In estimating the probabilities for escape, survival, and manifestation 
of adverse consequences of recombinant DNA molecules, one has to consider a 
number of factors. The probability of escape of recombinant DNA molecules 
from the laboratory depends on: one, the personnel, their training and use 
of acceptable procedures; two, the facilities, PI versus P2 versus P3 versus 
P4; three, the existence and implementation of an effective accident plan; 
four, the number of investigators using the techniques and the number of 
chimeric bacteria constructed and grown. The probability of survival and 
perpetuation of recombinant DNA molecules in nature depends on: one, the 
nature of the bacterial host and the cloning vector; two, the ecological 
niches occupiable by the original host and the genetically modified or 
chimeric host; three, the transmissibility of recombinant DNA to other 
bacteria; and four, the selective advantages or disadvantages conferred by 
foreign DNA. 
The probability of manifestation of adverse consequences of recombin- 
ant DNA molecules depends on: one, the source of the foreign DNA; and 
second, the expression of this DNA in a prokaryotic host. 
On the last slide (5) I have enumerated estimated probabilities for 
shotgun experiments with mammalian DNA using three different host-vector 
systems. These estimates are based on data whenever possible, and on 
[ 209 ] 
