FAQs

Opps, I cloned without --recursive

Fix with:

git submodule update --init --recursive
make all

The recursive flag (git clone --recursive ...) tells git to download additional git repos found within the McCortex repo. These repos are required libraries.

Why can't I edit the graph after adding links?

Links are a compressed representation of input sequence. Notably, they are compressed against the graph -- editing the graph results in the links not making sense anymore. Links record only the junction choices required to re-constitute the path through the graph.

Loading different combinations of samples along side each other does not count as editing the graph. Each sample's graph must not be edited after creating the sample's links.

See also: Infer edges

Coverage runs of less than k?

I'm assuming you've built a graph with some sequence data, and are aligned sequence data from a different sample.

A SNP in the aligned sample would usually give k zeros. However this is not always the case - here is an example of a 1 bp insertion into the sample aligned to the grap (k=5):

graph seq:  CCAAA-CC
sample seq: CCAAAACC
coverages:  1 0 0 1
kmers: CCAAA:1, CAAAA:0, AAAAC:0, AAACC:1

Why do I see very high coverage kmers?

High coverages are probably caused by repeats. They could also be due to contamination during sequencing. Repeats often mutate quickly, so a sample may have unique kmers either side of repeat kmers.