Hobbes FAQ

This FAQ provides answers to basic questions about Hobbes. You may also refer to the following papers for the technical details of Hobbes.

Questions

  1. Why doesn't Hobbes align any of the reads?
  2. What part of the code requires/uses popcnt instruction?
  3. How do I fix compilation errors?
  4. Why does Hobbes crash when building an index using hobbes-index?
  5. How do I find unalignable reads from Hobbes output.
  6. How were benchmark scores in the Hobbes2 paper generated and what do they mean?

Answers

  1. Why doesn't Hobbes align any of the reads?

    Given a distance threshold t, a read needs to contain at least t + 2 non-overlapping q-grams. That is, the length of a read needs to be greater than q * (t + 2). Otherwise, Hobbes doesnot align any of the reads.

    We recommend to set a gram length to "read length/(threshold + 3)" when building an index.

  2. What part of the code requires/uses popcnt instruction?

    Hobbes3 does not use popcnt instruction any more. Previous versions of Hobbes use popcnt to count the number of 1's in a BitVector. For the details of BitVectors, please refer to the following Hobbes paper.

    "Hobbes: optimized gram-based methods for efficient read alignment", Nucleic Acids Research 2011.

  3. How do I fix compilation errors?

    To complie Hobbes3, you need a proper version gcc (4.2 or later).

    If you choose to support compressed reads by running "build.sh compress", there might be link errors. In case of linker errors, you need to check names of shared libraries used in Hobbes3.

    The names of shared libraries of zip, bzip2, and boost iostream should be libz.so, libbz2.so, and libboost_iostream.so, respectively. However, the names of these libraries often contain version number in their suffixes. In those cases, you need to create a symbolic link. If gcc fails to find libboost_iostream in your system, for example, you need to make a symbolic link named "libboost_iostream.so" pointing to libboost_iostream.so.x.x.x, where x.x.x is a version number.

  4. Why does Hobbes crash when building an index using hobbes-index?

    Please check the genome file format. The problem is often caused by an additional blank line at the end of the file. Please delete additional blank lines at the end of a genome file.

  5. How do I find unalignable reads from Hobbes output.

    Hobbes3 outputs unaligned reads by default (Hobbes2 does not output unaligned reads as other mappers such as razerS3 and masai do not).

  6. How were benchmark scores in the Hobbes2 paper generated and what do they mean?

    They were generated from the Rabema benchmark. If you refer to the following Rabema paper, you can figure it out how to generate the values and what they mean.

    "A novel and well-defined benchmarking method for second generation read mapping", BMC Bioinformatics 2011.

2015 ISG | Website maintained by Jongik Kim | Created by Yun Huang | Original design Andreas Viklund