my recent attempt using redundans didn't work well

[SN]

Hello Leszek,

I’m using Redundans to help improve some de novo assembled genomes (assembled using Platanus with 250 bpPE+500bpPE+2kbMP+6kbMP+15kbMP libraries), I ran redundans on the scaffold.fa file generated by Platanus. 

My first two attempts – HN_red and HH_red generated good results, this was using the older version of Redundans in early 2017, I emailed you the log files. Redundans seemed to have worked very effectively in removing redundant scaffolds, though I don’t have the contigs.reduced.fa.hist.png file to look at the identity histogram. The reduced genomes had improved N50, expected genome size and reduced scaffold number.

My recent attempt was on a different assembly (Hmiss_redundans), and used same Platanus parameters as the above. After running redundans on the scaffold.fa, the resulting genome size was smaller than expected, a very big reduction in scaffold numbers and a good increase in N50. The contigs. reduced.fa.hist.png file (attached) shows that majority of scaffolds that were removed were not identical. Maybe I can fix it by increasing the –identity? But the other concern is the high variability seen in all the libraries which was not seen in my earlier attempts, I’m not sure if it is because of the newer version or the quality of the libraries was indeed bad. I emailed you both the Hmiss_redundans log file and the Hmiss_ reduced.fa.hist.png file.

I will appreciate any suggestions you could provide.

Thanks,

Sumitha

[l.p.p...@gmail.com]

Dear Sumitha, 

Thanks for interest in Redundans! 

I’m using Redundans to help improve some de novo assembled genomes (assembled using Platanus with 250 bpPE+500bpPE+2kbMP+6kbMP+15kbMP libraries), I ran redundans on the scaffold.fa file generated by Platanus. 

My first two attempts – HN_red and HH_red generated good results, this was using the older version of Redundans in early 2017, I emailed you the log files. Redundans seemed to have worked very effectively in removing redundant scaffolds, though I don’t have the contigs.reduced.fa.hist.png file to look at the identity histogram. The reduced genomes had improved N50, expected genome size and reduced scaffold number.

Great!  

My recent attempt was on a different assembly (Hmiss_redundans), and used same Platanus parameters as the above. After running redundans on the scaffold.fa, the resulting genome size was smaller than expected, a very big reduction in scaffold numbers and a good increase in N50. The contigs. reduced.fa.hist.png file (attached) shows that majority of scaffolds that were removed were not identical. Maybe I can fix it by increasing the –identity? But the other concern is the high variability seen in all the libraries which was not seen in my earlier attempts, I’m not sure if it is because of the newer version or the quality of the libraries was indeed bad. I emailed you both the Hmiss_redundans log file and the Hmiss_ reduced.fa.hist.png file.

If you suspect that the reduction is too severe, increase identity threshold. Unfortunately the attachments didn't arrived, so I cannot help you with that... 

Redundans complain about quality of libraries in two cases: 1. mate-pair library is contaminated with paired-end reads (then you see two types of read orientations, FR and RF) and 2. when insert size is highly variable. Sometimes it may happen that the library stats are poor due to high fragmentation of the initial assembly. If you believe that your libraries are ok, you can set the library stats manually as described in the manual.

Hope it helps, 

L.   

I will appreciate any suggestions you could provide.

Thanks,

Sumitha

--
You received this message because you are subscribed to the Google Groups "Redundans" group.
To unsubscribe from this group and stop receiving emails from it, send an email to redundans+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/redundans/fe157c93-cb5f-4c35-b9dc-deba80749117%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.