View Single Post
Old 02-01-2016, 12:16 PM   #3
moldach
Member
 
Location: Vienna

Join Date: Jan 2016
Posts: 10
Default More information

Hi Genomax,

I am working with a diploid eukaryotic transcriptome (not genome) from a coral species in the Acropora spp. complex.

There is another Acropora species which there is an available genome for [avg. sequence length ~1700bp; N50=~2200bp].

However, N50 is often misleading as it measures the continuity of contigs and not their accuracy; in transcriptome assembly the optimal contig is not known a priori and therefore carries little information . Similarly, for transcriptome assembly, these reference-free measures, as well as others (e.g. median contig length and number of contigs) can be misleading, or even meaningless, and should be avoided .

Therefore, I assessed the quality of my transcriptome assembly using Transrate; Transrate uses a reference genome/transcriptome to compare the quality of assembly. Because the A. digitifera genome is not annotated I used the annotated transcriptome of A. millepora. For my assembly, Transrate showed an initial score of 0.1316, and an optimized score of 0.2336 in Trinity. For comparison, approximately 50% of the de novo assemblies from the NCBI Transcriptome Shotgun Assembly database produce an overall score of 0.22 and optimized score of 0.35.

So my assembly is somewhat sub-optimal

Last edited by moldach; 02-01-2016 at 12:24 PM. Reason: punctuation
moldach is offline   Reply With Quote