Tal­mu­dic Ques­tion #27

How nar­rowly dis­trib­uted does a gene have to be to count as hav­ing no ho­mol­ogy with oth­ers? Even a brand-new, mys­te­ri­ously cre­ated cod­ing se­quence dropped into a sin­gle cell is go­ing to start spread­ing from there, ver­ti­cally and hor­i­zon­tally, and pretty soon it'll turn into a fam­ily of ho­mologs...
Vi­ral genomes seem to have a lot of genes with lit­tle or no ho­mol­ogy to genes from cel­lu­lar or­gan­isms, though. Where did they come from? (My fa­vorite ex­am­ple is Py­robac­u­lum spher­i­cal virus: a 28kb genome with 48 pu­ta­tive ORFS, none of which have any ho­mologs in the pub­lic se­quence data­bases — see Har­ing et al. 2004 Vi­rol­ogy 323(2): 233)

Dear En­rique and Ana,
that's an in­ter­est­ing ques­tion, the an­swer is an un­abigu­ous yes, be­cause of the for­mu­la­tion. Let's ask "...with all oth­ers?" And let's make it even more con­crete, leav­ing genes from outer space out of the is­sue for a minute. Let's con­sider the ori­gin of genes. IF we as­sume that the ori­gin of genes from the el­e­ments is pos­si­ble (which we are as­sum­ing here, I sup­pose, even for genes from outer space, un­less they some­how did not arise from the el­e­ments, but we ARE in outer space, so what, ex­actly, is the dif­fer­ence any­way...). Ahem. IF we as­sume that the ori­gin of genes from the el­e­ments is pos­si­ble, then does the ori­gin of some­thing that is al­most a gene but not yet a gene hap­pen only one sin­gle time — one mol­e­cule, from which all gene mol­e­cules de­scend? That is where your ques­tion leads, I think. Re­fus­ing to en­ter­tain the is­sue con­cern­ing the de­f­i­n­i­tion of a gene, I would muse that if there were mol­e­cules in a spot some­where that could give rise to one mol­e­cule of what one would call gene, then that the same process would be giv­ing give rise to lots of dif­fer­ent mol­e­cules of gene, if for no other rea­son than mol­e­cules tend to oc­cur in fairly large num­bers when the con­di­tions for them to oc­cur are right. Oth­er­wise we are all de­scended from one sin­gle pretty small mol­e­cule, which I have trou­ble find­ing at all likely.
Af­ter all, if there was only enough gene-like or­ga­nized C, N, O, H, P (and a dash of S in those pesky mod­i­fied tRNA bases) to make one sin­gle lit­tle bitty gene to be­gin with, then we have a pretty se­vere prob­lem mak­ing more of it for round two to get to luca or any­where else. That leads us to lots of mol­e­cules, more than one at the non-gene-to-gene thresh­old at least. And if bits of those dif­fer­ent first gene like mol­e­cules made some con­tri­bu­tion via their how­ever-repli­cated des­can­dants into what be­came a luca, then not all genes are ho­mol­o­gous (sim­i­lar by vir­ture of com­mon spe­cific mol­e­c­u­lar ori­gin in this ques­tion). So it would seem that even if all genes can be traced to luca, the an­swer is still not "easy", be­cause it de­pends upon how the genes in luca came to be, even if luca never stooped so low as to pocket a used gene from outer space. So small things in large num­bers con­sid­ered, the most sat­is­fy­ing an­swer ap­pears to be a com­fort­ing no.
Its like with cells: Genes al­ways arise from pre­ex­ist­ing genes, ex­cept at the ori­gin of genes from the el­e­ments, where the prob­lem sprouts new facets.