Dear Bastian, Alright so I used the recently released version of MIRA 4.0 to assemble a sample, and I had some PCR evidence for this sample. What I did, is that I amplified a locus, and cloned it, sequencing 32 clones. I am attaching a figure with obtained results. On the left I have the PCR clones that I sequenced, the number in parenthesis is the number of clones that had that specific haplotype. On the right I have 2 mira haplotypes produced from small contigs, smaller than 1.5kb. As you can see the second one is fine, it is the same 2nd one from the left figure. The first one made by mira however, seems to be a combination between #3, than #4 or #5 and than #1. It was not present in any of the PCR clones, so I decided to map reads to that first haplotype made by MIRA. I mapped with the following settings, 0% mismatch allowed, and only map paired reads, in the sense that for one read to map, it's pair needs to map nearby as well. As you can see on the bottom figure, the coverage sharply drops in the middle of the contig. In fact, in that area, only 1 read supports the link between position 5 and 6 in the haplotype (not seen in the bottom figure). No reads support the link between position 3 and 4. Would be very neat if MIRA could build variants and their respective frequencies for highly polymorphic loci. Thought I would share this. Adrian On Fri, Jan 31, 2014 at 3:58 PM, Bastien Chevreux <bach@xxxxxxxxxxxx> wrote: > On 31 Jan 2014, at 21:55 , Adrian Pelin <apelin20@xxxxxxxxx> wrote: > > [...] > > But within that one small contig, is everything phased? > > My lib is 100bp PE, insert about 300, and it sometimes builds small > contigs larger than 300. > > [...] > > Now as you can see, the first haplotype (GDR18) and the 3rd haplotype > (GDR19) are different potentially due to recombination. > > The first column corresponds to position 555 in a gene, and the 5th > column is position 791. So about 240bp distance between the 1st and 6th > column. The 4th column is position 649. The last column is position 959. > > > > Since these haplotypes represent small contigs, can I trust the way they > phase the variants? > > They very, very ... very probably are. There cannot be a 100% certainty > though. > > B. > > > -- > You have received this mail because you are subscribed to the mira_talk > mailing list. For information on how to subscribe or unsubscribe, please > visit http://www.chevreux.org/mira_mailinglists.html >
Attachment:
mira_vs_pcr.png
Description: PNG image
