Nanovar

* Fixed VCF header unclosed quotes (Thanks to Scott, https://github.com/cytham/nanovar/issues/5#issuecomment-592961341)
* Improved clustering algorithm (Clustering left and right breakends of each SV separately)
* Updated README.md

* Fixed cython compilation bug (Thanks to Scott, https://github.com/cytham/nanovar/issues/5#issuecomment-592766203)
* Updated requirements.txt

* Tool requirement changes:
* Added SAMTools
* Added Minimap2

* Algorithm and pipeline changes:
* NanoVar now uses both minimap2 and hs-blastn for increased sensitivity
* Upgraded model (v5)
* Improved time complexity for clustering algorithm
* Minor tuning of SV confidence scores by normal read coverage and number of alignments
* Default SV read support -c changed from 1 to 2
* Added progress information command-line feed during the run.

* Input/Ouput changes:
* NanoVar can now take a BAM file (tested BAM from minimap2) as input instead of FASTQ/FASTA reads
* Added ONT/PacBio option (For future use, currently both tech uses same model)
* Added custom model option
* Added --debug option
* Added Minimap2 executable path option
* Added SAMTools executable path option

* VCF format changes:
* "END" coordinate for BND SVs will now show POS+1 coordinate instead of "." to be usable for IGV (Thanks to Scott,
https://github.com/cytham/nanovar/issues/5#issue-559151039)

* Upgraded model (v4)
* Changed default score threshold to 1.0
* Added --mincov argument
* Added genotype prediction thresholds --homo --hetero
* Fixed make_interp_spline bug (Thanks to Asma, https://github.com/cytham/nanovar/issues/1#issue-536971382)
* Fixed logging oversize
* Fixed HTML figures and file path link
* Disabled sorting by bitscore

* Fixed svread-overlap.tsv file formating in nv_nn.py
* Fixed operation bug in nv_valid.py l201

* Python 3.5 is no longer supported