what are your thouths about this result

[nicedad]

Hello all,

I need your suggestion about this result I got from a de novo transcriptome analysis using Newbler to gain best results (I think by improving N5).
I changed min. overlap length each time as in the table below and have the following quesitons:

What do you think, will this have an impact on the analysis?.
Which criterion we may use to gain best result for further analysis (i.e. # of contigs, N50....ect).

Mi.Overlap Length    #duplicate   N50   Total contigs      mean   Total seq.
35                         106     912    15056           604.8   9106339
40                         136     904    15417           592.5   9134740
45                         132     899    15487           590.8   9152710
50                         113     893    15537           589.3   9155956

cheers,

[westerman]

I am not sure of your question. It looks likes the minimum overlap of 35 bases gives the best results -- fewer duplicates, higher N50, lower number of contigs, better mean. The total number of bases is less but this metric is neither good nor bad unless you know the number of bases that you should be getting.

I would hesitate to go below 35 bases as an overlap because you could start getting false positives but you could give it a try.

[nicedad]

Thanks westerman,
However, I performed the analysis with min. overlap of 30 and this did change anything and written down the total number of bases.
Do you think the 35 is the best value of the min. overlap?

Mi.Overlap Length    #duplicate   N50   Total contigs      mean    Total bases
30                    106         912    15056             604.8   9106339           
35                    106         912    15056             604.8   9106339    
40                    136         904    15417             592.5   9134740    
45                    132         899    15487             590.8   9152710    
50                    113         893    15537             589.3   9155956