NOVEL METHOD FOR GENERATING AND SCREENING AN ANTIBODY LIBRARY The present invention relates to the field of monoclonal antibodies, among others human antibodies. More particularly, the present invention is directed to a novel method for generating a DNA sequence coding for a heavy chain and/or a light chain of at least one antibody, preferably from total or messenger RNA from a cell capable of expressing an antibody. The invention also comprises a method for generating a bank or library of antibodies, preferentially human antibodies, applying the steps of the previous method. The expressions «bank(s)» or «library(ies)» will be indifferently used in the present description, both of these expressions having the same meaning. Traditionally, monoclonal antibodies or fragments thereof were prepared by using the standard technique of hybridomas described by Kohler and Milstein (Kohler and Milstein, 1975, Nature 256, 495), the thereby obtained antibodies being then humanized. More recently, novel methods, based on the understanding of natural mechanisms and on the development of DNA recombination techniques, have been developed for generating and expressing human monoclonal antibodies. In particular, the techniques aiming at obtaining combinatorial libraries, or banks of antibodies have been the subject of many developments. Such techniques in addition to the fact of being more rapid and with which it is possible to do without humanization steps, provide a more efficient use of the whole and of the variability of the list of antibodies. Antibodies are glycoproteins of the superfamily of immunoglobulins formed with 4 polypeptide chains: 2 heavy chains (H for heavy) and 2 light chains (L for light) which are connected to each other by a variable number of disulfide bridges providing flexibility to the molecule. These chains form a Y-shaped structure and consist of immunoglobulin domains of about 110 amino acids. Each light chain consists of a constant region and of a variable region while the heavy chains consist of a variable region and of 3 or 4 constant regions depending on the isotype. For a given antibody, the two heavy chains are identical, the same applies for the two light chains. The constant regions are characterized by amino acid sequences which are very closely related from one antibody to the other, and are characteristics of the species and of the isotype. Each light chain has one exemplary thereof noted as CL. The heavy chains include three or four constant regions CH1, CH2, CH3 and CH4 depending on the isotype. An antibody has four variable regions located at the ends of the two «arms». The association between a variable region borne by a heavy chain (VH) and the adjacent variable region borne by a light chain (VL) forms the recognition site (or paratope) of the antigen. Thus, an immunoglobulin molecule has two sites for binding to the antigen, one at the end of each arm. Both of these sites are identical, hence the possibility of binding two antigen molecules per antibody. For heavy chains, the variable regions VH are coded by genes which include three types of segments, i.e. a variability segment (v segment or gene), a diversity segment (d segment or gene) and a junction segment (j segment or gene). The genes coding for the variable regions of the light chain VL also have a gene v and a gene j, but no gene d. Initially, the naive immunity repertoire is derived from the combinatorial rearrangement of different genes coding for the variable regions. More particularly, the genes v, d, j for the heavy chain and v, j for the light chain, may be associated with each other independently, which explains the large diversity of the variable regions which may be generated. This first mechanism by which a very large number of functional units of different structures may be produced, is called combinatorial diversity. Two other mechanisms will increase this diversity, one occurring in the initial step of differentiation of B lymphocytes (junctional diversity) and the other one during the immune response induced by antigenic stimulation (somatic hypermutations). Junctional diversity is based on the appearance of mutations during the combinatorial rearrangement at junctions between the genes v-d and d-j for the heavy chain and v-j for the light chain. A maturation of the affinity of the antibodies is carried out by a process for selecting somatic hypermutations which consist in point-like mutations exclusively occurring in the regions coding for the regions VH and VL and which may be the origin of changes in the structure of the paratope. Diversity is therefore based on the three described elements, i.e. combinatorial diversity, junctional diversity and somatic hypermutations. Starting with these elements, different technologies have been developed in order to be able to rapidly generate a large diversity of human antibodies and have proved to be particularly interesting in the case of human antibodies for which the hybridomas are technically difficult to obtain. More particularly, the advent of the so-called «phage display» technology (Smith, 1985, Science, 228:1315; Scott & Smith, 1990 Science, 249:386; McCafferty, 1990, Nature, 348:552) has given the possibility of selecting in vitro, from libraries, antibodies or fragments of human antibodies directed against a large variety of targets, or antigens. These libraries are divided into two groups, i.e. natural libraries and synthetic libraries. Natural libraries are built from genes directly recovered from human B cells, therefore after different mechanisms ensuring combinatorial diversity, junctional diversity and somatic hypermutatoins. With such libraries, it is therefore possible to naturally generate a large variability of antibodies. Further, these natural libraries have the advantage of using an established and validated technology and allow compatibility with various expression systems such as bacteria, yeasts, eukaryotic cells or plants. However, generation of these libraries remains unwieldy and difficult to apply and requires resorting to a PCR (Polymerase Chain Reaction) step. For applying this PCR step, it is necessary to know the respective sequences of the 5' and 3' regions, in order to define from known germinal lines, the primers required for this PCR. If this point is not particularly bothersome for the end 3' which is located at a gene c (i.e. coding for the constant region of the antibody), it is particularly limiting as regards the end 5' which, as for it, is located at the end of the gene v, i.e. of the variability gene. This required knowledge of the region 5' of the gene v, inherently limits the selectable primers to the primers from germinal lines known up to now and therefore to the known gene v sequences. This has the effect of limiting the antibodies which may be recovered, to the antibodies having known sequences at the region 5' of their variable region, therefore setting aside all the other antibodies for which the sequences of the ends 5' of the genes v would not be known. Further, it may be necessary to modify the amino acids in the variable region in order to facilitate cloning and/or amplification. Another drawback resulting from the required knowledge of the end 5' of the gene v, is the risk of missing the antibodies for which somatic hypermutations would have occurred at this end. Finally, because of the PCR technique, only the antibodies for which the end 5' of the gene v will have very strong affinity with the primers used, the lower affinities not being retained, will be recovered as a priority. The synthetic libraries, as for them, are built from different non-rearranged genes. Their rearrangement is performed in vitro by PCR and subsequently, different in vivo maturation mechanisms may be applied in order to obtain the best antibodies in terms of specificity and affinity. By resorting to synthetic libraries, it is possible to increase the diversity of the obtained antibodies and this remains compatible with the various expression systems. However, nonetheless it is true that the work required for generating them, remains long and laborious and many drawbacks persist. Indeed, for producing synthetic libraries, a PCR step always remains mandatory. In addition to the drawbacks mentioned above, related to the use of PCR, this step requires the use of large size degenerated primers which frequently introduce deletions, additions and modifications of base pairs within the assembled v, d, j genes, which may also result in the formation of antibodies incapable of folding themselves properly and therefore non-functional. Further, the antibodies selected from these libraries are generally weakly expressed and in many cases contain mutations which may affect the immunogenicity of the antibody. Further, with the generation of synthetic libraries, it is not possible to take into account the mutations at the junction areas ensuring junctional diversity of the somatic hypermutations. There is therefore a large loss of variability. The present invention aims at overcoming the whole of the drawbacks listed above by describing for the first time a simple method for generating a DNA sequence coding for a heavy chain and/or a light chain, and most particularly for generating a library of antibodies having large diversity. The object of the invention is thus a method for generating a natural library of heavy chains and/or light chains of antibodies not only having its advantages but also those of a synthetic library, while getting rid of the respective drawbacks of each of the two technologies. More particularly, the present invention differs from the whole of the techniques known up to now by doing without the PCR amplification step and consequently the whole of the drawbacks described above related to this PCR step. Most particularly, the technology, object of the present invention, no longer requires that the sequence of the end 5' of the gene v of the variable domain of the antibody be known. The present invention differs from the prior art by the amplification step which is carried out on a single strand DNA sequence. Indeed, unlike the whole of the prejudices of one skilled in the art and the state of the art limited to a standard PCR amplification step, i.e. on a double strand DNA molecule, the present invention describes for the first time a method for generating a bank of antibodies from a single primary primer and from the single strand cDNA sequence obtained from this primary primer. According to a first aspect, the invention, object of the present patent application, is directed to a method for generating a DNA sequence coding for a heavy chain and/or a light chain of at least one antibody from an RNA extract or mixture from a cell, this cell being capable of expressing RNA coding for an antibody or at least for the heavy chain and/or the light chain of an antibody, characterized in that said method comprises at least the following steps: a) putting a primary primer (I) in contact with said RNA extract or mixture liable to contain at least one RNA, designated here as a «sense» RNA, coding for the heavy chain or the light chain of an antibody, this primary primer being capable of specifically hybridizing to a fragment of the sequence of said «sense» RNA, comprised in the sequence coding for the constant region C of the heavy chain and/or light chain of said antibody; b) synthesizing from said primary primer (I), a single strand cDNA designated as «anti-sense» cDNA; c) eliminating if necessary, the primary primers (I) which have not hybridized in step a); d) annealing said «anti-sense» cDNA by introducing a covalent bond between its end 5' and its end 3'; e) putting a secondary primer (II), designated as «sense» secondary primer, in contact with said «anti-sense» circular cDNA obtained in step d), this «sense» secondary primer (II) being capable of hybridizing to said «anti-sense» circular cDNA; f) amplifying said «anti-sense» cDNA from said secondary primer (II); and g) recovering the thereby amplified «sense» complementary linear DNA strand. Resorting to a PCR step requires knowledge of the sequences of the ends 5' and 3' respectively of the gene which has to be amplified in order to be able to hybridize complementary primers therein and to thereby initiate amplification by conventional techniques of the double strand molecules. The actual principle of the invention is based on the capability of amplifying an annealed single strand DNA sequence with a suitable enzyme. It therefore clearly appears to one skilled in the art that only one complementary primer of the end 3' of the sequence which should be amplified is required. More particularly, within the scope of generating a sequence coding for an antibody, only a complementary primer of all or part of the sequence corresponding to the constant region will then become necessary for the amplification according to the invention. Once the primer is attached on the single strand DNA molecule, in the present case, the «sense» RNA molecule, it is sufficient to amplify the latter with a suitable enzyme capable of amplifying a single strand sequence. One skilled in the art will easily understand that the advantage brought by the invention, object of the invention, in the sense that the latter allows entire amplification of the «sense» RNA sequence corresponding to the variable region, and this as far as its end 5', i.e. including if necessary, the sequences corresponding to the signal peptide as well as to the region 5'-UTR. This aspect is particularly advantageous, in the sense that the antibody isolated by this technique may then be directly cloned with its original signal peptide, which is optimum for expressing such antibodies in a eukaryotic expression system. Indeed, without the presence of the original signal peptide, cloning of antibodies in an expression vector requires adding of a non-native signal peptide, which always requires knowledge of the sequence of the end 5' of the gene of the antibody and may introduce modifications of nucleic and/or protein sequences of such antibodies. Further, with such a method, it is possible to get rid of the whole of the drawbacks described above. It is quite obvious that each step described above may be replaced with an equivalent step with which the essential result sought for this step may be obtained without however departing from the object of the present invention and from the desired scope of protection for this method. For the sake of clarity, it is specified here that, in the whole of the present patent application, the expressions «sense» and «anti-sense» have as system of reference, in a standard way for one skilled in the art, the 5' 3' direction of the initial RNA molecule, i.e. the RNA molecule coding for the heavy chain or the light chain of an antibody. Accordingly, and logically, the primary primer (I) and the cDNA molecule generated from this primer, both being complementary of said RNA, are therefore characterized by the «anti-sense» expression. In the same way, the secondary primer (II) as well as the linear DNA molecule generated from this primer, both complementary of the generated cDNA and therefore oriented like the initial RNA sequence, will be characterized by the expression «sense». Finally, in the same way, the tertiary primer (III) and the complementary DNA sequence generated form this tertiary primer will, as for them, be characterized by the expression «anti-sense». According to a preferred embodiment of the invention, the cells from which stem the extracts or mixtures of RNA are mammalian cells, preferably from mice, rats, or primates, the human origin being most preferred. In a preferred embodiment, the RNA extract or mixture is an extract or mixture of total RNAs, an extract or mixture of RNAs enriched with messenger and/or pre-messenger RNA. According to the second embodiment, the method, object of the invention, may be directly applied to natural, healthy cells, or cells from a pathological tissue or organ, among others a tumor, these cells being capable of expressing RNAs coding for a heavy or light chain of an antibody. The method, object of the invention, may also be applied to cells of a cell line from healthy or pathological cells. According to an also preferred embodiment of the invention, the extracts or mixtures of RNAs originate from splenocytes, nodules or B lymphocytes. Preferably, the cells from which stem the extracts of mixtures of RNA are B lymphocytes or one of its progenitors or a plasmocyte. Preferably, these cells are selected from pro-B progenitor cells (stem cells already engaged in B lineage), pre-B precursor cells, immature B lymphocytes, immunocompetent mature B lymphocytes or plasmocytes (see among others for the characteristics of these cells, the definitions given in the publication «Differenciation des lymphocytes B» by Professors Marie-Paule LEFRANC and Gerard LEFRANC, Universite Montpellier II and Laboratoire d'ImmunoGenetique Moleculaire, LIGM, on the ImmunoGenetics site