STAR outputs interpretation

Question

[XBria]

Hi everyone,

I am working on Rna-seq data. Star is mapping only on chromosome X, data are down-sampled. That is why the uniquely mapped alignments rate is close to 100. (paired-end , length of forward75, reverse,75)

Can I say the mapping is improved using these parameters ?

--outFilterMatchNmin 20 --seedSearchStartLmax 30 --outFilterScoreMinOverLread 0 --outFilterMatchNminOverLread 0 --outFilterMismatchNoverLmax 9

* I examine 3 samples out of 12, all three tests emerge the exact value of 98.24 rate of uniquely mapped reads !!!!!

Following is the output without these parameters:

 Started job on |   Dec 06 09:32:03
                         Started mapping on |   Dec 06 09:32:11
                                Finished on |   Dec 06 09:33:52
   Mapping speed, Million of reads per hour |   47.10

                      Number of input reads |   1321477
                  Average input read length |   152
                                UNIQUE READS:
               Uniquely mapped reads number |   1281329
                    Uniquely mapped reads % |   96.96%
                      Average mapped length |   151.01
                   Number of splices: Total |   701323
        Number of splices: Annotated (sjdb) |   693399
                   Number of splices: GT/AG |   697684
                   Number of splices: GC/AG |   1968
                   Number of splices: AT/AC |   703
           Number of splices: Non-canonical |   968
                  Mismatch rate per base, % |   0.43%
                     Deletion rate per base |   0.01%
                    Deletion average length |   1.53
                    Insertion rate per base |   0.01%
                   Insertion average length |   1.29
                         MULTI-MAPPING READS:
    Number of reads mapped to multiple loci |   16029
         % of reads mapped to multiple loci |   1.21%
    Number of reads mapped to too many loci |   286
         % of reads mapped to too many loci |   0.02%
                              UNMAPPED READS:
   % of reads unmapped: too many mismatches |   0.00%
             % of reads unmapped: too short |   1.80%
                 % of reads unmapped: other |   0.01%
                              CHIMERIC READS:
                   Number of chimeric reads |   0
                        % of chimeric reads |   0.00%

and with those parameters:

  Started job on |  Dec 06 09:45:39
                         Started mapping on |   Dec 06 09:45:43
                                Finished on |   Dec 06 09:48:04
   Mapping speed, Million of reads per hour |   33.74

                      Number of input reads |   1321477
                  Average input read length |   152
                                UNIQUE READS:
               Uniquely mapped reads number |   1298240
                    Uniquely mapped reads % |   98.24%
                      Average mapped length |   150.22
                   Number of splices: Total |   705133
        Number of splices: Annotated (sjdb) |   696785
                   Number of splices: GT/AG |   701438
                   Number of splices: GC/AG |   1991
                   Number of splices: AT/AC |   710
           Number of splices: Non-canonical |   994
                  Mismatch rate per base, % |   0.45%
                     Deletion rate per base |   0.01%
                    Deletion average length |   1.51
                    Insertion rate per base |   0.01%
                   Insertion average length |   1.29
                         MULTI-MAPPING READS:
    Number of reads mapped to multiple loci |   22727
         % of reads mapped to multiple loci |   1.72%
    Number of reads mapped to too many loci |   406
         % of reads mapped to too many loci |   0.03%
                              UNMAPPED READS:
   % of reads unmapped: too many mismatches |   0.00%
             % of reads unmapped: too short |   0.00%
                 % of reads unmapped: other |   0.01%
                              CHIMERIC READS:
                   Number of chimeric reads |   0
                        % of chimeric reads |   0.00%

[WouterDeCoster]

Okay you increase the fraction of mapped reads, but how do you know those alignments are also "correct"? The percentage aligned, although informative, shouldn't be your only parameter for optimization.

I also wouldn't bother about a difference of only 2%.

[GenoMax]

Ref: Improving the mapping rate by aligner parameters

XBria : Creating new threads with variations of the questions from before is not going to help much. At the high end of alignment %, you are splitting hairs. Like @Wouter said above, if those 2 additional % are adding meaningfully to your analysis is questionable. You should be moving on to the actual differential expression analysis.

[XBria]

Dear Genomax, Could you share a link that clearly explains the trade-offs, I could not find any comprehensive resources on this issues. I need to know more about this. Thanks for understanding me as a beginner in this scope

[GenoMax]

What trade-offs are you referring to?

[XBria]

I mean the best choices among different optimization results. (it may result in withdrawal of some parameters to be set) How to clearly recognize if we are heading an optimal way ? based on which criteria ?

[XBria]

The aim is to improve mapping through parameters setting. How may I know if they are correct ? Is it not showing a good improvements then, is that right ?