How to improve the mapping rate of vg giraffe

Question

[gulin]

Hello! I am using vg giraffe to map the short read data(150bp) against HPRC pangenome graph. The mapping rate is about 92% in my test data with a quite low perfectly mapping rate (25%). However, using bwa-mem to map the same data against CHM13 linear reference results in a better mapping rate(98%). To my opinion, mapping to pangenome can reduce the reference bias in mapping. Moreover, the results in MC paper can also prove that giraffe is better than bwa-mem. How can I adjust the parameters to improve the mapping rate? Or how can I prove using giraffe and pangenome reference is better than bwa-mem and linear reference?

Answer 1

[colindaven]

You'll get a lower mapping rate, but a higher number of perfect alignments. That is the improvement. You should get less heterozygous SNPs downstream.

Look here for example at around 38.5 minutes -