FIELD OF THE INVENTION The present invention relates generally to the field of medicine, and specifically to microbiology, immunology, vaccines and the prevention of infection by a bacterial pathogen by immunization. BACKGROUND OF THE INVENTION Streptococcus pneumoniae is a leading cause of meningitis, pneumonia, and severe invasive disease in infants and young children throughout the world. The multivalent pneumococcal polysaccharide vaccines have been licensed for many years and have proved valuable in preventing pneumococcal disease in elderly adults and high-risk patients. However, infants and young children respond poorly to most pneumococcal polysaccharides. The 7-valent pneumococcal conjugate vaccine (7vPnC, Prevnar®) was the first of its kind demonstrated to be highly immunogenic and effective against invasive disease and otitis media in infants and young children. This vaccine is now approved in many countries around the world. Prevnar contains the capsular polysaccharides from serotypes 4, 6B, 9V, 14, 180, 19F and 23F, each conjugated to a carrier protein designated CRMigy. Prevnar covers approximately 80-90%, 60-80%, and 40-80% of invasive pneumococcal disease (IPD) in the US, Europe, and other regions of the world, respectively [1,2]. Surveillance data gathered in the years following Prevnar's introduction has clearly demonstrated a reduction of invasive pneumococcal disease in US infants as expected (FIG. 1) [3,4]. Surveillance of IPD conducted in US infants prior to the introduction of Prevnar demonstrated that a significant portion of disease due to serogroups 6 and 19 was due to the 6A (approximately one-third) and 19A (approximately one-fourth) serotypes [5,6]. Pneumococcal invasive disease surveillance conducted in the US after licensure of Prevnar suggests that a large burden of disease is still attributable to serotypes 6A and 19A (FIG. 1) [3]. Moreover, these two serotypes account for more cases of invasive disease than serotypes 1, 3, 5, and 7F combined (8.2 vs. 3.3 cases/100,000 children 2 years and under). In addition, serotypes 6A and 19A are 1 associated with high rates of antibiotic resistance (FIG. 2) [7,8,9]. While it is possible that serogroup cross-protection will result in a decline of serotype 6A and 19A disease as nnore children are immunized, there is evidence to suggest that there will be a limit to the decline, and a significant burden of disease due to these serotypes will remain (see below). Given the relative burden and importance of invasive pneumococcal disease due to serotypes 1, 3, 5, 6A, 7F, and 19A, adding these serotypes to the Prevnar formulation would increase coverage for invasive disease to >90% in the US and Europe, and as high as 70%-80% in Asia and Latin America. This vaccine would significantly expand coverage beyond that of Prevnar, and provide coverage for 6A and 19A that is not dependent on the limitations of serogroup cross-protection. SUMMARY OF THE INVENTION Accordingly, the present invention provides generally a multivalent immunogenic composition comprising 13 distinct polysaccharide-protein conjugates, wherein each of the conjugates contains a capsular polysaccharide from a different serotype of Streptococcus pneumoniae conjugated to a carrier protein, together with a physiologically acceptable vehicle. Optionally, an adjuvant, such as an aluminum-based adjuvant, is included in the formulation. More specifically, the present invention provides a 13-valent pneumococcal conjugate (13vPnC) composition comprising the seven serotypes in the 7vPnC vaccine (4, 6B, 9V, 14, 18C, 19F and 23F) plus six additional serotypes (1, 3, 5, 6A, 7F and 19A). The present invention also provides a multivalent immunogenic composition, wherein the capsular polysaccharides are from serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F and 23F of Streptococcus pneumoniae and the carrier protein is CRM197. The present invention further provides a multivalent immunogenic composition, wherein the capsular polysaccharides are from serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9v, 14, 18C, 19A, 19F and 23F of Streptococcus pneumoniae, the carrier protein is CRM197, and the adjuvant is an aluminum-based adjuvant, such as 2 aluminum phosphate, aluminum sulfate and aluminum hydroxide. In a particular embodiment of the invention, the adjuvant is aluminum phosphate. The present invention also provides a multivalent immunogenic composition, comprising poiysaccharide-protein conjugates together with a physiologically acceptable vehicle, wherein each of the conjugates comprises a capsular polysaccharide from a different serotype of Streptococcus pneumoniae conjugated to a carrier protein, and the capsular polysaccharides are prepared from serotype 3 and at least one additional serotype. In one embodiment of this multivalent immunogenic composition, the additional serotype is selected from the group consisting of serotypes 1, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F, and 23F. In another embodiment, the carrier protein is CRM197. In yet another embodiment, the composition comprises an adjuvant, such as an aluminum-based adjuvant selected from aluminum phosphate, aluminum sulfate and aluminum hydroxide. In a particular embodiment, the adjuvant is aluminum phosphate. The present invention also provides a multivalent immunogenic composition, comprising poiysaccharide-protein conjugates together with a physiologically acceptable vehicle, wherein each of the conjugates comprises a capsular polysaccharide from a different serotype of Streptococcus pneumoniae conjugated to a carrier protein, and the capsular polysaccharides are prepared from serotypes 4, 6B, 9V, 14,18C, 19F, 23F and at least one additional serotype. In one embodiment of this multivalent immunogenic composition, the additional serotype is selected from the group consisting of serotypes 1, 3, 5, 6A, 7F, and 19A. In another embodiment, the carrier protein is CRMigy. In yet another embodiment, the composition comprises an adjuvant, such as an aluminum-based adjuvant selected from aluminum phosphate, aluminum sulfate and aluminum hydroxide. In a particular embodiment, the adjuvant is aluminum phosphate. 3 The present invention also provides a method of inducing an immune response to a Streptococcus pneumoniae capsular polysaccharide conjugate, comprising administering to a human an immunologically effective amount of any of the immunogenic compositions just described. The present invention further provides that any of the immunogenic compositions administered is a single 0.5 mL dose formulated to contain: 2 pg of each saccharide, except for 6B at 4 pg; approximately 29 pg CRM197 carrier protein; 0.125 mg of elemental aluminum (0.5 mg aluminum phosphate) adjuvant; and sodium chloride and sodium succinate buffer as excipients. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 depicts the changes in IPD rates by serotype in US children <2 years of age from baseline (1998/1999) to 2001. FIG. 2 depicts the distribution of pneumococcal isolates with resistance to penicillin (PCN) in children <5 years of age (1998). FIG. 3 depicts the reverse cumulative distribution curves (RCDC) of OPA post-third dose results from the 0118-P16 Prevnar trial. DETAILED DESCRIPTION OF THE INVENTION Inclusion of Prevnar Serotypes 4, 6B, 9V, 14, 18C, 19F, 23F Data from IPD surveillance between 1995-1998 estimated that the seven serotypes in Prevnar were responsible for around 82% of IPD in children <2 years of age [5]. In Northern California, the site of the efficacy trial, the Prevnar serotypes accounted for 90% of all cases of IPD in infants and young children [10]. Since introduction of the Prevnar vaccine in 2000, there has been a significant decrease in the overall IPD rates due to a decrease in disease due to the vaccine serotypes [3,4]. Therefore, there is no justification at this time to remove any of the Prevnar serotypes from the next generation of pneumococcal conjugate vaccines but rather to add serotypes to obtain wider coverage. 4 Inclusion of Serotypes 1, 3, 5 and 7F In the US, the rate of IPD caused by serotype 1 in children under the age of 5 years is <2%, about the same as for each of types 3 and 7F [1,6]. Serotypes 1 and 5 account for higher rates of IPD in US populations at high risk for invasive pneumococcal disease. Specifically, serotype 1 causes 3.5% of IPD in Alaskan native children <2 years of age, and 18% in children 2-4 years of age [11]. Both serotype 1 and serotype 5 significantly cause disease in other parts of the world and in indigenous populations in developed countries [12,13,14]. Serotype 1 may also be associated with more severe disease as compared with other pneumococcal serotypes [15]. This observation is based on the difference in rates of case identification between the US and Europe, and the associated difference in medical practice. Overall, the incidence of IPD is lower in Europe than in the US. However, the percent of IPD caused by serotype 1 in Europe is disproportionately higher than in the US (6-7%, vs. 1-2%, respectively). In Europe, blood cultures are obtained predominantly from hospitalized children. In the US, it is routine medical practice to obtain blood cultures in an outpatient setting from children presenting with fever >39°C and elevated white blood cell counts. Given the difference in medical practice, it is postulated that the lower percent of disease caused by serotype 1 in the US may be diluted by higher rates of other serotypes causing milder disease, while the higher percent in Europe reflects more serious disease. In addition, seroepidemiology studies of children with complicated pneumonia demonstrate that serotype 1 is disproportionately represented [16,17,18]. This suggests that inclusion of serotype 1 may reduce the amount of severe pneumococcal disease, as well as, contribute to a total reduction in invasive pneumococcal disease. The addition of serotypes 3 and 7F will increase coverage against IPD in most areas of the world by approximately 3%-7%, and in Asia by around 9%. Thus, an 11-valent vaccine would cover 50% in Asia and around 80% of IPD in all other regions [1,2]. These serotypes are also important with respect to otitis media coverage [19]. In a multinational study of pneumococcal serotypes causing otitis media, Hausdorff et al found serotype 3 to be the 8th most common middle ear fluid 5 isolate overall [20]. Serotype 3 accounted for up to 8.7% of pneumococcal serotypes associated with otitis media. Thus, the importance of types 3 and 7F in otitis media, as well as in IPD, warrants their inclusion in a pneumococcal conjugate vaccine. However, attempts to produce a multivalent pneumococcal conjugate vaccine that exhibits significant immunogenicity with respect to serotype 3 polysaccharides have been unsuccessful. For example, in a study of the immunogenicity and safety of an 11-valent pneumococcal protein D conjugate vaccine (11-Pn-PD), no priming effect was observed for serotype 3 in infants who had received three doses of the vaccine followed by a booster dose of either the same vaccine or a pneumococcal polysaccharide vaccine (Nurkka et al. (2004) Peof. Inf. Dis. J., 23:1008-1014). In another study, opsonophagocytic assay (OPA) results from infants who had received doses of 11-Pn-PD failed to show antibody responses for serotype 3 at levels comparable to other tested serotypes (Gatchalian et al., 17"^ Annual Meeting of the Eur. Soc. Paed. inf. Dis. (ESPID), Poster No. 4, PI A Poster Session 1, Istanbul Turkey, Mar. 27, 2001). In yet another study, which assessed the efficacy of an 11-Pn-PD in the prevention of acute otitis media, the vaccine did not provide protection against episodes caused by serotype 3 (Prymula et al. www.thelancet.com. Vol. 367: 740-748 (March 4, 2006)). Accordingly, a pneumococcal conjugate vaccine comprising capsular polysaccharides from serotype 3 and capable of eliciting an immunogenic response to serotype 3 polysaccharides provides a significant improvement over the existing state of the art. Inclusion of Serotypes 6A and 19A a. Epidemiology of Serotypes 6A and 19A Surveillance data in the literature suggest that serotypes 6A and 19A account for more invasive pneumococcal disease in US children <2 years of age than serotypes 1, 3, 5, and 7F combined (FIG. 1) [1,5]. In addition, these serotypes are commonly associated with antibiotic resistance (FIG. 2) and play an important role in otitis media [6,19,20]. The ability of the current Prevnar vaccine to protect against disease due to 6A and 19A is not clear. The rationale for inclusion of 6A and 19A components in a 13vPnC vaccine is discussed below. 6 b. Responses to 6A and 19A Induced by 6B and 19F Polysaccharides The licensed unconjugated pneumococcal polysaccharide vaccines (for use in persons at least two years of age) have contained 6A or 6B capsular polysaccharide but not both [21]. Immunogenicity data generated at the time of formulation of the 23-valent pneumococcal polysaccharide vaccine demonstrated that a 6B monovaient vaccine induced antibody to both the 6A and 6B capsules. The data from several trials assessing IgG and opsonophagocytic assay (OPA) responses in a variety of populations with free polysaccharide and with pneumococcal conjugate vaccines suggested that IgG responses to 6A are induced by 6B antigens, but the responses are generally lower, and the OPA activity with 6A organisms is different than with 6B organisms [22,23,24,25]. In addition, subjects responding with high 68 antibody may have little or no activity against 6A. In contrast to the chemical composition of the 6A and 6B capsular polysaccharides where there exists a high degree of similarity, the 19A and 19F capsules are quite different due to the presence of two additional side chains in the 19A polysaccharide. Not surprisingly, immune responses measured in human volunteers immunized with 19F polysaccharide vaccine showed that responses to 19F were induced in 80% of subjects, but only 20% of subjects had a response to 19A [26]. Low levels of cross-reactive IgG and OPA responses to serotype 19A after immunization with 19F polysaccharide have also been documented in trials with conjugate vaccines as well [24,26]. Internal data on cross-reactive OPA responses to 6A and 19A have been generated from the 7vPnG bridging trial (D118-P16) conducted in US infants (FIG. 3). These studies are consistent with the findings of others, and demonstrate induction of cross-reactive functional antibody to 6A polysaccharide after immunization with 6B polysaccharide, although at a lower level, and very little functional antibody to 19A after immunization with 19F. 7 Impact of 6B and 19F Immunization on 6A and 19A in Animal Models Animal models have been used to evaluate the potential for cross-protection with polysaccharide immunization. In an otitis media model developed by Giebink et al., chinchillas were immunized with a tetravalent polysaccharide outer membrane protein (OMP) conjugate vaccine (containing 6B, 14, 19F, 23F saccharides) or placebo [27]. In this trial there appeared to be some cross-protection for 6A; however this did not reacii statistical significance and tfie level of protection was lower than with 6B against otitis media. In this same model there was 100% protection against 19F otitis media, but only 17% protection against 19A otitis media. Saeland et al. used sera from infants immunized with an 8-valent pneumococcal tetanus conjugate vaccine (containing 6B and 19F) to passively immunize mice prior to an intranasal challenge with 6A organisms, in a lung infection model [28]. Of the 59 serum samples, 53% protected mice against bacteremia with 68 and 37% protected against 6A. Mice passively immunized with sera from infants immunized with four doses of an 11-valent pneumococcal conjugate vaccine (containing 19F conjugated to tetanus toxoid) were given an intranasal challenge with 19A organisms in the same model [29]. Of 100 mice passively immunized and then challenged, 60 mice had no 19A organisms detected in lung tissue, whereas organisms were identified in all mice given saline placebo. However, passive immunization did not protect against challenge with 19F organisms in this model; therefore, the relevance of the model for serogroup 19 is questionable. In general these models provide evidence of some biological impact of 68 immunization on 6A organisms although the effect on the heterologous serotype was not as great as that observed with the homologous serotype. The impact of 19F immunization on 19A organisms is not well understood from these models. Impact of 6B and 19F Polysaccharide Conjugate Immunization on 6A and 19A Disease in Efficacy/Effectiveness Trials The number of cases of disease due to the 68, 6A, 19F and 19A serotypes in 7vPnC and 9vPnC (7vPnC plus serotypes 1 and 5) efficacy trials is noted in Table 1 [30,10,31]. The numbers of invasive disease cases are too small to allow any conclusions to be drawn for serotypes 6A and 19A. However, the Finnish otitis 8 media trial generated a large number of pneumococcal isolates [32]. In the per protocol analysis 7vPnC was 84% (95% CI 62%, 93%) efficacious against otitis media due to serotype 6B and 57% (95% CI 24%, 76%) efficacious against otitis media due to serotype 6A (Table 1). In contrast, serotype-specific efficacy with the 7vPnC was not demonstrated for otitis media due to either 19F or 19A. Table 1. Cases of Pneumococcal Disease Due to Serotypes 6B, 6A, 19F, and 19A in Efficacy Trials with the 7vPnC and 9vPnC Vaccines ] 6B I 6A I 19F I 19A PnC IContr. PnC IContr. PnC iContr. PnC IContr. Kaiser Efficacy Trial - 7vPnC A 7 n A 2*13 0 1 (ITT) Navajo Efficacy Trial - 7vPnC 05101 110 South African Efficacy Trial -A OI n n ^ T 1 9vPnC HIV (-) (ITT) I z i u u i o i South African Efficacy Trial -A -, q-ino i •> ■» 9vPnC HIV (f) (ITT) ^ / d 1U ^ 1 dose) vs. unvaccinated, and adjusted for underlying conditions Reference 40 and personal/confidential communication A published analysis [3] of the use of Prevnar also indicated that serotypes 6B and 19F conferred a moderate reduction in IPD caused by serotypes 6A and 19A among children under two years of age (Table 1 in [3]). Disease rates among unimmunized adults caused by serotypes 6A, 9A, 9L, 9N, 18A, 18B, 18F, 19A, 19B, 19C, 23A and 238 ("all vaccine-related serotypes") were somewhat reduced (Table 2 in [3]). These data establish that herd immunity from the use of Prevnar in children under two years of age was modest for serotypes 6A and 19A, and provide a basis for the inclusion of serotypes 6A and 19A in the 13vPnC vaccine of this invention. Conclusion for addition of 6A and 19A The post-marketing surveillance data and the case-control study results noted in FIG. 1 and Table 2 with the TvPnC vaccine suggest that, consistent with the other information on immune responses and performance in the animals models described above, there may be some cross-protection against 6A disease, but to a lesser extent than to 68 disease. Furthermore, it appears the protection against 19A is limited. Therefore, a 13vPnC vaccine containing serotypes 6A and 19A provides coverage that is not dependent on the limitations of serogroup cross-protection by serotypes 68 and 19F. 10 Accordingly, the present invention provides a multivalent immunogenic composition comprising 13 distinct polysaccharide-protein conjugates, wtierein each of the conjugates contains a different capsular polysaccharide conjugated to a carrier protein, and wherein the capsular polysaccharides are prepared from serotypes 1, 3, 4, 5, 6A, 6B, 7F, 9V, 14, 18C, 19A, 19F and 23F of Streptococcus pneumoniae, together with a physiologically acceptable vehicle. One such carrier protein is the diphtheria toxoid designated CRM197. The immunogenic composition may further comprise an adjuvant, such as an aluminum-based adjuvant, such as aluminum phosphate, aluminum sulfate and aluminum hydroxide. Capsular polysaccharides are prepared by standard techniques known to those sl CD o> CO" -A CO g -f in rt S ^^c\j'>-cMooooora 3 D5 S V-'r-CM-<-CO-!-rtT-T--*CnCMv-g J 3 ^ c ?7cNrixro?'2^^fj^s~cNrjf!oS"rj^ .=; COCON.05:rfST-CDi^h~CJ)CD^ 5 b ^ ?? CD '"- N. *^-- ^ '"- cj, ° >« ^" oo '*- - S- O ^- CO °'.. cv, ^- g CD- in q. „ °> S O c ji^J S CM t in in cvi S ^ V-5 g CO o> P oD't g CO CO S S CO CD 00 53'^ s O ^^sSa'S-<^«-S<=>^"feS"^«J^'«"coS-^-"e?5°o^"ciS4 ? U. a)">£"^in"-^r^-<-co'"inc\iS'"cD'*ffi-<--cn'"->-csjf^:IZ'<--<-'t'"CT> •- ..^Js'^ i-~ t^ o ■* "-"-in °lr- ^3- f^. ^in CO co-g S3 "-^ '~' ■— p _, ,« 'c 2 "* s> 5 ^^ O CD « 2 SS»?° O O ■* -'t O CM ^ 'S- O O O O^S ^ o 3% ^ t- -"-^ d d T^ d CO d i-^ -.-^ T-^ T^SS = o§ 5- E 5 < -* °>g e3j:>:2 2 ° S ^ O en in f^ CD O O o-oX — C '5 •«- -^ T- 1- T- T- „ ^ 'JZ T- f- -^ T-mCJ QQ;>5V V V V V v'"'-""^ V V V V'wg ^,4,4 C on r- 0) o CL .*i - « roii in Table 4. S. pneumoniae OPA GMTs for NZW Rabbit Serum Pools Following Immunization with Two Doses of 13-vaient Pneumococcal Glycoconjugate -— l3vPnC^ ~13vPnC-t-ALPO? WeekO Week 4 wk4-vui,n WeekO Week 4 Serotype WK4.WkO Wk4:WkO 1 <8 64 16 <8 64 16 3 <8 8 2 <8 16 4 4 <8 16 4 <8 32 8 5 <8 128 32 <8 512 128 6A 8 128 16 8 512 64 6B <8 256 64 8 1,024 128 7F 8 64 8 8 128 16 9V 8 64 8 8 128 16 14 16 32 2 16 32 2 18C 8 256 32 <8 256 64 19A <8 256 64 <8 1,024 256 19F <8 128 32 <8 512 128 23F 8 64 8 <8 256 64 A: Pools consisted of equal volumes of serum from individual rabbils within a treatment group {n=12) Study #HT01-0036 Study #HT01-0036 compared rabbit immune responses to the polysaccharides (PSs) contained in the vaccine, after immunization with the 13vPnC vaccine with or without conjugation to the CRM197 protein. New Zealand White rabbits were immunized intramuscularly at weel< 0 and week 2 with a dose of 2.2 |ig of each PS (except 4.4 \xg of 6B). Animals received one of three vaccine preparations: (a) 13vPnC {PS directly conjugated to CRMigy), (b) 13vPnPS, (free PS) or (c) ISvPnPS + CRM197 (free PS mixed with CRM197). All vaccine preparations contained AIPO4 as the adjuvant at 125 ^g/dose. Serotype specific immune responses for all vaccine preparations were evaluated in an IgG ELISA and complement-mediated OPA measuring functional antibody. The immune responses were compared between the treatment groups. Table 5 presents GMT data obtained from week 4 bleeds analyzed in antigen specific IgG ELlSAs. Additional analyses show the ratio of GMT values at week 4 to week 0. The data indicate that the conjugate vaccine preparation elicited greater serum IgG titers than free PS or free PS + CRM197 vaccine. With the exception of S. 46 pneumoniae type 14, the 13vPnC vaccine was able to induce functional antibodies to tlie representative strains of S. pneumoniae in an OPA (Table 6). After two immunizations with either the 13vPnPS or 13vPnPS + CRM197 vaccine, neither could induce OPA titers > 8-fold at week 4 relative to week 0 for 10 out of the 13 serotypes measured (Table 6). In conclusion, these results indicate that conjugation of the 13-valent pneumococcal vaccine polysaccharides produces higher serum IgG titers and overall greater functional antibody activity than seen with free polysaccharide alone or mixed with unconjugated CRM197. 47 9 T- fV. O ^ CM ^. °^. '^. f^ ^. °. O '^. '* '^. ^. '^ O ■* 7? O ^" ■>* CO O 'H ^ „- ;V ^ a> nC CN! cvi" CO CM T- (^J X- ^ «2 C ^ 8> in o o 2: csT rL U5 f^ ^ gJ CO ^ «7 g o g o fe -" to M [S of « in ^ w cn o" Xr ■^" N-" (N •*'" '^" CM" f^- CM" CM" m" S CL^^£i'°£iCi.£2.ii-cocz::,£i O C- ^ a i •^ SJ^inmin^^inwin^mcDCD ■s s «= s ~ ■SSooo'T^^P^ocn'^cD^co « i5'Siri-^^cDCM£::iT-^cocoi:T-:2iri E ^ 3 — SE O) S K .—. --^ ■—- .—. .—, .—. E rPo 03.-^a>--~-,-^o,^^-^_.in,_^T- 5 ^iS "g'^.„^-'°-«'^- U- a.— coco-— oo^^^^^^cococo ^ tn -£ _j "^ Ui >» o ■°JJ "Sm'^OOOOCMOinocDOT-X'. C/)(5 SSooioincncoininincoixicDE^ DL O) 5 ^ C 3 Q--E' o O O n -^ ■♦"•o -Scocvfcoino'^cM-r-in'^oo'^^^ ^ CD o_ of ? ia> 'o^fcoSint^oococno-c-co (0'