In general I will try to:

1. Mapping the reads against genome
2. Variant calling
3. Variant filtering
4. Extract those variants falling within your gene region (if you know your gene, you can extract the coordinates using bed/gtf/gff3 file)
5. Create consensus (there are different methods and tools, like GATK, vcf-consensus, bcftools consensus... )

My reads come from another organism, and the sequence of gene from another. I followed some steps using BWA, like aligning paired read ends using bwa aln, then combining using bwa sampe, then I used samtools to create a bam file and samtools mpileup to create a consensus.