Various improvements and bug fix

Bug fix: fix repeated adding samples to CountReads.samples during loading so that the mean reference counts for all samples can be calculated accurately. This would slightly affect the parameter minmean in count_analysis and thus affect the inclusion of low-read genes in the analysis.

Add pathway and GO enrichment analysis and pathway drawing

1. Dynamically download KEGG pathway data based on biopython module Bio.KEGG.

2. Add enrichment analysis with p-values calculated by Bayesian method to achieve more robust statistics (using freqcdf).

3. In side freqcdf, algorithm using log beta function betaln replaced direct beta function for numeric stability when calculating p-values for large samples.

4. Include function getGeneId() to download the newest mapping of gene symbol/Ensembl ID to Entrez gene id using mygene.MyGeneInfo().

5. Include function draw_kegg_pathways() to draw the changes onto KEGG pathway images for better visualization of what has changed. The code is modified from a code I found online several years ago, unfortunately I can no longer find the source to give appropriate acknowledges.

PhD thesis submission

Updated neutralstats.py to handle errors for linalg.lstsq on older versions of python3.