I am having trouble compiling my program when I use htslib. Since this is software many other people are likely to use, I imagine my pain * 10000. Is there a way to read bam files without htslib? I just need the chrom, start, end and strand.
Otherwise, is there a way to include all of htslib in my software?
I am using the bioconda htslib and the flag -Ihtslib when compiling, but I get the error:
fatal error: htslib/sam.h: No such file or directory
#include "htslib/sam.h"
^~~~~~~~~~~~~~
•
written
by
Click downvote
-Ihtslib means there is a directory named htslib in your current directory and this directory a bunch of *.h files.
if this directory is not in the current working directory , search for it (someting like)
find / -type -f -name "sam.h" 2> /dev/null
(wait if you have a large server). Then use
-I/path/to/the/directory/htsdlib
I am using the bioconda htslib
does it contains the sources of htslib ? try to install htslib from github without conda
So I can include the whole htslib-project in my project if I want other users to compile easily?
So I can include the whole htslib-project in my project
what does 'include the whole htslib-project in my project ' means ? are you talking about git submodules ? https://git-scm.com/book/en/v2/Git-Tools-Submodules
I want people to not need to have htslib installed when they compile my software. So I can have a folder with htslib in my codebase and point to it.
You can: It is MIT-licensed. https://github.com/samtools/htslib/blob/develop/LICENSE
HTSlib's headers are set up so that they are in an htslib/ subdirectory of an includedir and used with #include "htslib/sam.h" etc. This avoids collisions with other libraries that might provide unrelated headers coincidentally named sam.h.
So if your find command shows /path/to/the/directory/htslib/sam.h, the -I option you should use will be -I/path/to/the/directory.
My solution. Works, but probably not the cleverest. Hinges on users having conda w/appropriate libs installed:
# In setup.py
from subprocess import check_output
conda_path = check_output("which conda", shell=True).decode().strip()
conda_include = []
conda_lib = []
if conda_path:
conda_base = conda_path.replace("bin/conda", "")
conda_include.append(conda_base + "include/")
conda_lib.append(conda_base + "lib/")
extensions = [Extension("bla.src.reads_to_bins",
["bla/src/reads_to_bins.pyx", "bla/src/gzstream.cpp"], language="c++",
include_dirs=conda_include,
library_dirs=conda_lib,
extra_compile_args=compile_options,
libraries=["z"]),
Chapter 7 of Computer Systems: A Programmer's Perspective was of great help. The book is highly recommended for people who want to understand computers from a C programmers perspective.
R version:
library(GenomicAlignments)
## import bam
gr <- readGAlignments('./your_indexed_bam_file.bam')
## converts each bam entry into colon separated strings (e.g. "chr2:11109-11282:-")
## it may help to convert to a GRanges object in a separate step if your bam is huge
as.character(GRanges(gr))
Log in to answer this question.
