tag:blogger.com,1999:blog-8081751738815878503.post6931740378115673382..comments2023-04-14T03:40:29.342-04:00Comments on Sui Generis Brewing: Identifying Wild Yeast Part DeuxBryanhttp://www.blogger.com/profile/16672407110077541595noreply@blogger.comBlogger5125tag:blogger.com,1999:blog-8081751738815878503.post-14273731931755151252013-05-14T08:13:57.266-04:002013-05-14T08:13:57.266-04:00RE: Pichia. The high rate of "false positive...RE: Pichia. The high rate of "false positives" I get for pichia is (IMO) exactly for the reason you list - brett & pichia are very close, and may even intermix.<br /><br />That there is a high degree of relatedness is apparent in my samples - the matches between Pichia (WLP653) and Meyerozyma/Pichia (WLP650) are perfect matches in the constant region of the 26S rRNA. Given the closeness of the species, the perfect match in the constant region is what we'd expect.<br /><br />But there are also clearly differences - if you blast the full WLP650 or 653 sequences, despite getting 100% matches with Meyerozyma/Pichia, there's consistently ~300bp of the input sequence that is cut off of the results. This "missing" section is the D1-D3 variable region (basepairs 1-300 in the 653 example, 600-end in the 650). If you BLAST the D1-D3 regions (and nothing else) you get the 'correct' match.<br /><br />I believe that the WLP strains are from a brewery, but they may be in a strain library as well.Bryanhttps://www.blogger.com/profile/16672407110077541595noreply@blogger.comtag:blogger.com,1999:blog-8081751738815878503.post-16762404212708431792013-05-13T17:47:27.830-04:002013-05-13T17:47:27.830-04:00I would for sure link to your site. I mean, it is ...I would for sure link to your site. I mean, it is your data and I don't want to get credit for that. In addition, I always mention my sources whenever possible. I just like to ask first before using other's data/results.<br /><br />I would like to compare some phylogeny trees and compare them with the shift assay. It seems logical to me that one underestimates divergences on protein level. I am kind of interested now to see if differences are present in phylogeny trees (based on 26S rDNA) and the shift assay. <br /><br />If I recall correctly, the sequencing results based on the ITSs gave you some Pichia hits as well as for the 26S rDNA. In Yeasts, chapter 13, Kurtzman et al mention the conundrum of assigning Dekkera to a different species (or one of its own). Interestingly, D. bruxellensis appears in the Pichia clade if one constructs a Ascomycetous yeast tree. However, in phylogeny trees with Pichia and other Pichia-close related species (and Dekkera), Dekkera appears as a distant clade. What if Dekkera (or some of the strains) are hybrids of yeast species X and Picha and therefore show similarities to Picha sp. 26S rDNA (because the 26S rDNA is from the Pichia species)? <br /><br />Would be interesting to know if the two White Labs Brettanomyces strains originate from a yeast database (and therefore would have a identifier like CBS XX). <br /><br />I guess there is a very interesting answer why the two White Labs strains show such a divergent 26S rRNA sequence (and ITSs). Cheers, SamAnonymousnoreply@blogger.comtag:blogger.com,1999:blog-8081751738815878503.post-41944289096471109092013-05-13T16:16:17.784-04:002013-05-13T16:16:17.784-04:00oops. Where I wrote 'driven by the hypersensi...oops. Where I wrote 'driven by the hypersensitive of longer/partial matches to score higher than short/perfect matches.' <br /><br />I meant<br /><br />"driven by the perpensity of the alignment algorithm to score higher for longer/partial matches than short/perfect matches."Bryanhttps://www.blogger.com/profile/16672407110077541595noreply@blogger.comtag:blogger.com,1999:blog-8081751738815878503.post-17806450425531865402013-05-13T16:14:11.941-04:002013-05-13T16:14:11.941-04:00Sam, everything here is open to the public; no nee...Sam, everything here is open to the public; no need to ask permission to use anything (a link-back would be nice though...). I'll try to be brief in my answers.<br /><br />The DNA isolation method I used is crude, so I cannot spec the DNA and get an accurate reading. I suspect there is little-to-no DNA in the samples that didn't work. As I say at the end of this post, I'm planning on trying another method that avoid the DNA isolation all together.<br /><br />All my analysis was done using NCBI nucleotide blast, with the mentioned species filters, aside from the cladogram.<br /><br />For my tree I trimmed the sequences down to the D1-D3 variable regions. Otherwise you get a massive bias towards the constant region and "false" alignments with things like picha.<br /><br />As for your tree, it perfectly reflects the issues I was having - if you don't take into account the varying sizes of the rRNA sequences in the databases, you get unusual clustering of the sequencing results driven by the hypersensitive of longer/partial matches to score higher than short/perfect matches. It is possible to 'tweak' the clustering algorithms to more strongly select against mismatches, but I'm not familier enough to do that properly. Instead I limited my trees only to the D1-D3 variable region (including the non-variable bits between the D1/D2 and D2/D3). The resulting region is ~300bp in size.<br /><br />You are correct that electrophoretic shift assays are no longer used - but keep in mind that they tend to underestimate evolutionary divergence. Synonymous mutations, and mutations which substitute amino acids with similar chemical properties (i.e. small hydrophobic for small hydrophobic) tend to be missed.Bryanhttps://www.blogger.com/profile/16672407110077541595noreply@blogger.comtag:blogger.com,1999:blog-8081751738815878503.post-54277927508018061182013-05-13T15:13:24.437-04:002013-05-13T15:13:24.437-04:00(posted by eurekabrewing.wordpress.com; I can'...(posted by eurekabrewing.wordpress.com; I can't post with wordpress somehow)<br />Bryan, thanks for sharing your results. Interesting results indeed. I had to do some in-silico analysis with your data myself. May I use your two WLP yeast sequences for a future post of mine? Would make it easier to talk about some of my results as well. I have some other questions. <br /><br />- Troubleshooting the failed amplification for the Saccharomyces strains. Did you use the same amount of DNA for all the four PCRs? (Measured the DNA concentration). <br /><br />- What seq alignment algorithm did you use to compare the sequences (to get the pairwise identities for B. naardenensis)?<br /><br />- Have you trimmed your sequences before building the phylogeny tree? I aligned your sequences a 26S rRNA multiple sequence alignment and trimmed the overhangs in your sequences. Just to get better multiple seq alignments for better/faster phylogeny trees later on. <br /><br />- Have a look at one of my trees... (http://www.ebi.ac.uk/Tools/services/web_clustalw2_phylogeny/toolresult.ebi?tool=clustalw2_phylogeny&jobId=clustalw2_phylogeny-I20130513-193853-0176-18004711-pg) What are your thoughts about that?<br /><br />- Concerning the electrophoretic shift assay. The 5th edition does not show the table from above anymore. Simply because this method has been criticized since it is based on proteins rather than DNA. Like to cite Kurtzman et al (The Yeasts, a Taxonomic Study, 5th edition): "1. The method assays the genotype only indirectly, so that much variation at the nucleotide level may go undetected because nucleotide substitutions do not necessarily change the amino<br />acid composition; 2. Changes in amino acid composition do not necessarily change the electrophoretic mobility of the protein and, as a consequence, alleles that are considered to be the same protein alleles from different individuals may represent different gene<br />alleles"<br /><br />I haven't compared all my phylogeny trees with the mobility table above. I therefore don't want to make any conclusions yet if the (DNA based) phylogeny trees represent the mobility phylogeny relationships. <br /><br />Cheers, Sam<br />Anonymousnoreply@blogger.com