Hi! I am trying to build a consensus contig using assembly from a few reads that satisfy certain criteria. I extract reads from a bam file and then check through my scripts if they meet the criteria to select a small number of them. I have typically 10 to 12 reads. These reads are on an average 5 kbp long.
The reads are from a bam file that has been created using bwa-mem on PacBio data for human.
I used MIRA for the local assembly, but unfortunately I do not see any outputs in a .maf file. In fact, no .maf file is present under the _test_Assembly_d_results directory.
While MIRA itself completes just fine (without any fatal errors) I do get this "critical warning" in the log file. (Of course, the DNA I am assembling is not bacterial: see text below.)
Is there a way to choose parameters in MIRA that can help with this problem?
Any help is appreciated. Thanks!
-------- CRITICAL warning --------
MIRA warncode: CONCOV_SUSPICIOUS_DISTRIBUTION Title: Suspicious distribution of contig coverages
- 0 contig(s) with a total of 0 bases (= -nan% of bases in all non-repetitive large contigs) have an average coverage less than 75% of the average coverage of all non-repetitive large contigs.
- 0 contig(s) with a total of 0 bases (= -nan% of bases in all non-repetitive contigs) have an average coverage more than 125% of the average coverage of all non-repetitive large contigs.
- 0 contig(s) with a total of 0 bases (= -nan% of bases in all non-repetitive contigs) have an average coverage 25% above or below the average coverage of all non-repetitive large contigs. Summary: found 3 indicator(s) for coverage problem(s).
If the DNA you are assembling is bacterial, this could indicate that you sampled and sequenced DNA from exponential or late exponential phase of a bacterial population. This leads to a coverage bias toward the origin of replication, hence false positive detection of repeats, hence an assembly which is more fragmented than it could be or may have misassemblies in regions located toward the opposite of the origin of replication. Only available countermeasure: for your next sequencing project, do not sample in exponential phase but sample in stationary phase (if possible).
