Twentieth, Dear Lederberg, _ Z was very glad to receive your letter, and . _ to have the opportunity of a more leisurely examination _ of your data. sO se “ NAN RR EI Miah nt ta AA RAE RAE IE ALESIS NE atone First of all I checked the agreement of the “soupling” and “repulsion” phases in regard to recombination of Ty and the marker genes in the sets of data numbered 1, 2 and 3. in your letter. They agrees well in 2 and 3 but not so well in 1. Sugh occasional disagreements are not, however, uncommon in ordinary linkage data and this one should therefore probably not, be taken too seriously. = | -, .. . Secondly I: cheoked. the agreement of the reoom- me ad bination value for T, and the marker genes in sets 1,2, ~ : 3 and.4, pooling the eoupling and repulsion data erm "2 ..8@t8 1, 2 and 3. Sets 1 and 3 agree in sh “and 16.1% recombination, sets’ 2 and 4 agree wii . and 27.3%. 1 and-3'on the one hand and 2-and 4 on the other clearly disagree. Since { ani 3 have TP while 2 and 4 have TL, (we ean exclude By from consideration. _ ane ‘on the evidence of ‘set 4)we must assume that Ty is. >” | linked to L and P, rather than to T, the Giftferente:in i. . Eecombination being due to the differmoe in positden - Sy of and P. T might af oouree be in the same chromesome, ._ buf it, must therl'be further from 7, than'l, and P are. °~ nose are not brought: into this a eement, on the denve of set 4. This is confirmed by the lack of effect of changing from B, @, C to B, M on the reoomb- ination value, though it should be noted that 4f% 9° vere onierly inked with B the same result would. be 3 ooserved «. a : - dnéependent of one another in the data whether the De J. Lederberg =... 5. Contdnusd....... 20th January ,1947 | é Data set 4 indicates the order you Bive, viz. s but we do not know whether this should not really be somthing like > ‘\ where reflects not true Linkage but the association of two unlinked genes enforced b; your teohnique, whieh can of gourse lead to genes in ferent chromosomes appearing a6 if in the seme branebed chromosome. By and Lae would both show Linkage with BM and yet appear arrangement was“ 2 ss or of the kind... f This ambiguity oan be removed by separating B and M~ in the exseriusn ts. The recombination value of By with . B dade is direotly caloulable as’ or about 9% ae shy veeombination values of a oan be ecatimted. if we care to assume the absence of - interference. . Setting the recombination values at ~ Dr J. Lederberg ..... Continued .......-20th January ,1947 we have your four classes ~1ry -S4 +v, + s vropertional bo 41 P2 43 » 44 Go Py » Py ap a3 and By Po Py respectively q= i-p. We can then find py = 10% , po = 26% and By = 16% These add up to 58% but, as you will see, the accuracy of the estimate depends largely on the frequency of the rare + s triple crossover class. i remember that you thought your frequency of 5 for this class to pe too high, in wnich case the values of py --- D3 would be over-estimated. ag you say, all this deyends on the assumption of linear arrangement except, of course, in so far as we recoznise the possibility of = branched appearance being spuriously engendered by the teohnique of insisting tha Bu etc. be recovered tozether. Such genes may or may not be linked. Your remark that types such as mt, te tit ere rare suggests linsage of B and M eto., but if they all turn out to he linked the result would be somewhat surpris- ing. It may be that linkage of the type we know is not operating in your bacterium, but I think we must at this Stege see how far our present ideas of linkage can explain your results. it also seems to me that the data on strongly Suggest @ linear order, as the + s class occurs With about the right frequency on such a view. We might expect it otherwise to be higher. A more ordered tes of linear order could be made with (a) 3 linked genes none of which was used as a marker (o) 2 linked genes also linked to a marker and both "outside" the rezion between the markers, 140. not between BM and LT. In this case care mist be taken to exclude the possibility of an arrangement such as which would obviously give a branched appearance. Pree eee en . United States. Dr J. Lederberg ......... Continued ........ 20th January ,1947 (c) (I think) 3 gene between the two murkers iee. Seain situation ef the tyves woulda have to be excluded. Possibility (ce) requires & bit more examination but I think & linexr order could be tested with its aid. I have not yet had an opportunity to discuss the matter with Fisher, so that all tne above is fase my own opinion at present. I hope that you find it (a) intelligible, (b) useful and (¢) sound. I find the problem & very intrigu- ing one and I hope that I shall see some more of it, or better still, have the opportunity of another personal discussion with you as enjoyable as the one in New Haven. In any case I am sure that you are right in exploring all the possibilities of explaining the results on standard linkage theory as a first step. Please give my beat wishes to Mrs Lederberg. I hope (or should I say, expect 7) that you are both enjoying your new status. Yours sincerely, Dr J. Lederberg, Osborn Botanioal Laboratory, Yale University, New Haven, Connecticut, Fn ea ne ene ere are ogg!