Chris Miller

Moderator approval of the first post seems reasonable to me, as long as it can trigger an email notification or something. I'm happy to approve some posts if it means cutting down on this nonsense. If Istvan's new countermeasures don't work out, we could give it a shot. - Chris Miller

Right. This calculation used single-end reads, so the numbers will be lower than what you can get from paired-end reads, using that extra information. - Chris Miller

By doing self-alignment, we can see exactly what percentage of the genome is uniquely mappable with different size reads. I've got these results laying around: To be clear these are calculated by taking each possible read of a given length, mapping it back with BWA, and then determining whether it is mapped uniquely to the correct position. You can calculate these yourself in a pretty straightforward manner, or look at the "Mapability" track in UCSC to grab some pre-computed ones. - Chris Miller

Breakdancer doesn't call copy number variants. It calls structural variants. If you see a large deletion in breakdancer, it may be the case that a copy number deletion has occurred. It's also possible that the deleted sequence got re-integrated somewhere else, in which case, there would be no effect on copy number. - Chris Miller

It may depend on the source, but in my experience, it's defined as the percentage of bases that are incorrectly called. The 0.8% error rate that you describe would mean that of every 1000 bases coming off the sequencer, 8 of them will report the incorrect base. - Chris Miller

I'm not sure I entirely understand your question, but there doesn't appear to be anything wrong with the output you posted. It indicates that you have 7 mutations in LATS1, all SNVs. If you had 2 indels and 2 SNVs, the total mutations would still be 7. According to the FDR values, this gene is mutated more frequently than would be expected, based on the background mutation rate in the samples. - Chris Miller

There isn't any single answer to this question. It all depends on what kind of biological question you're trying to answer with the RNAseq data. If I'm looking for differential exon usage due to spliceosome mutations, gene-level data is useless to me. If I'm trying to work with a huge network of genes, I may need to simplify my inputs and use gene-level metrics to make the problem tractable. - Chris Miller

No worries, man. If I had a dollar for every time something turned out to be way harder than I expected... - Chris Miller

a) I don't know of any off-the shelf tools, b) this is generally difficult, because breakpoint-spanning reads may map to one side or the other, but also may not map at all. To get a high confidence call, you could create a short contig containing your breakpoint sequence +/- 200 bp, append it to the reference genome, then realign all of your reads against this. Compare the depth of breakpoint spanning reads on your contig to the depth at the breaks on the original reference sequence, and that ratio will give you a pretty good idea of the frequency. - Chris Miller

Even if it was a missing zero typo, 3200 a month would be 38,400 a year, which is less than a postdoc salary. - Chris Miller

I'm assuming POA = "Plan of Attack". You're going to have to be a lot more specific about what you're asking in order to get any help here. I'm closing this post. If you have a specific question about some aspect of gene expression analysis, feel free to try again with a new question. - Chris Miller

1) You can use an orthogonal method (another algorithm) 2) For real validation of structural variants, we generally do capture, then deep sequencing of the breakpoint regions, then use an assembly-based process to confirm the presence of reads that support the breakpoint - Chris Miller

I imagine you're using a newer version than was used in the original paper. I believe that there have been many updates/improvements to the algorithm over the past few years. - Chris Miller

Agreed. You can either include the "chr" prefix or not, but be consistent across all of your files. - Chris Miller

A MAF file typically consists only of variants predicted to be somatic. This means that they should be present in the tumor, but not the normal (ignoring for the moment low-level tumor contamination in the normal). These tumor-specific variants should be annotated in the MAF file, according to the specifications here: https://wiki.nci.nih.gov/display... - Chris Miller

Depends on your samples. If this was an extremely inbred mouse strain, then you might be able to say with some confidence that the reference strain has an AC at this position, so observing only A is likely to be due to skewed expression and not a mutation. In humans, you might also be able to look at allele frequencies from dbSNP or 1000 genomes to get an idea of how often you see A and C in the population. To be completely sure that you're seeing an RNA-level skew, though, you really need to know what the alleles are in the DNA. - Chris Miller

No, that's not normal. My guess is that one of your steps failed partway through and only processed the first portion of the genome. Check your err logs. - Chris Miller

You can't detect allele-specific expression in RNA-seq unless you know the allelic status of the DNA. For example, if you see entirely As expressed at one location, does it mean that the DNA is entirely A, or does it mean that it's a het A/C with allele-specific expression? You could, however, detect a subset of biased expression patterns. If you see 90% As and 10% Cs, you can be fairly confident that the DNA is A/C and that you're seeing biased expression of the A allele. The caveat here is that you'll miss any site that is 100% expression of one allele or the other (for the reasons outlined above). Really, to do this type of analysis, you're going to want to have RNA and DNA sequence from the same sample. - Chris Miller

Yep, Sean nailed it. SNVs that only appear on one strand or the other are often false-positives. That said, if sequence coverage is low at that site, you may only see reads from one strand just by chance. It's up to you to decide what acceptable thresholds are. - Chris Miller

Yes, this is true. - Chris Miller