Entering edit mode
9.5 years ago
SJ Basu
▴
50
Hello,
I am working with miRNA for quite some time now and use miRDeep2 for it. I have these two queries to which i've searched a lot for the answer but without any success.
- The read depths for miRBase hits (both
result.csv&expression.csv) doesn't equal to total reads of the input file. I suppose it should be like Total reads= miRBase hits + novel miRNAs isn't it? Or am I going wrong - what is the last part of
result.csvsignifies?
My pipeline: raw reads-> cutadapt -> rRNA filtering -> mapping to genome -> miRDeep2
Any help in this regard will be highly appreciated. Thanks in advance!
What happens when you add in the unmapped reads?
Sir,
I am not exactly sure which unmapped reads you are pointing to cause I am performing mapping twice (with rRNA then genome). I do create the .arf file and
reads_collapsed.fafrom the sam file (which has both mapped and unmapped reads) of reads mapped to genome usingbwa_sam_converter.pl(provided by miRDeep2). Now this arf automatically takes in only the mapped reads from the sam file.Hi,
I don't know exactly the reason, but I guess it should be Total = miRBase hits + novel miRNA + everything else. You are guessing that everything not in miRBase is novel miRNA, something that is not true. There are a bunch of sequences that are not miRNAs, like tRNA, or fragments of genes, or repeat regions.
Thanks a lot for affirming my belief :) Lorena I knew about this fact but what's I am upset about is I am providing 7.1 million reads of which 2.3 million reads is in 17-22bp region but miRDeep2 is showing only 194346 reads(miRBase + novel). Don't you think its too less? Actually me and a senior of mine we were running to different programs and confirming the sure miRBase hits (he is running miRNAkey). Now is the most disturbing part when I ran miRNAkey on the same data, he got 138 significant miRNA and I got 78 significant miRNA! I am clueless about where the mistake is happening
whoo...that's a lot. That only happens when the species is wrong. About the number of DE miRNAs, don't know if you are using the same tool to that or not. But if not, that is normal.
I don't know if you are saying that you both got the same number of hits, or you got much less that your colleague?? if so, it seems you have something else there, maybe other species miRNA. If not, something happened with miRBase.
I will happy to run one sample with another tool if you only tell me the species, just out of curiosity.
okay so miRDeep2 and miRNAkey outputs may not be same - I got that. But the problem is when my colleague is running miRNAkey its giving 100000+ reads in 138 miRBase hits (significant ones that is p value < 0.05) and when I am running miRNAkey on the same data in my computer its giving 8000+ reads in 78 miRBase hits (significant ones that is p value < 0.05)! I updated miRBase through update option in the miRNAkey GUI. I got this version of miRBase http://mirbase.org/ftp.shtml and
grep ">osa"to separate out my species then indexed it. Oh yes my species is Oryza sativa.Well, if you are using exactly the same, and get different results....that's weird. Some reason could be tools version, and parameters. Or your input files for some reason are different. It can not be more than that....this is weird.
I did check my parameters and tool version, they are fine and matching. I am kind of in a fix with this problem! Though, me and my senior we were discussing the fact that may be our filtration step is giving different number of reads to the input file. Meanwhile I tried miRPlant and output looks pretty satisfying. Can you provide me with some opinion on this tool?
I read about that, not my expertise because I never did plant miRNA, but it seems good. If you are getting better results, just inspect them, If they make sense, all good. Sorry to not help a lot in this case. It seems that the parameters to play most are the one related to the hairpin size, or min loop length, but I think your species is not so weird, so I will play not much, maybe some different values to see difference in results.
ok. I shall do so and lets see how else can I solve my original problem though you were really helpful. And thank you so very much for all time you took out for this post. :)
Hello
Recently I am analyzing sRNA data..and I saw your post because I am also exploring that which tool is better for this analysis. If I am not wrong then 194346 reads from collapse file that's why it is showing less number.
Hello,
No Deepika actually I did the calculation from collapsed file too. The thing is they enlist only mapped miRNAs, and mirdeep2 is good for animals but kind of falls short with plants so, I went with miRPlant and it was pretty satisfying. If you are analyzing plants try that or else miRDeep2 is good.