Add ANI field

Dockerfile

@@ -1,21 +1,20 @@
-FROM ubuntu:16.04
-MAINTAINER Alexey Gurevich "https://github.com/alexeigurevich"
+FROM ubuntu:24.04
+LABEL maintainer="Alexey Gurevich" maintainer_url="https://github.com/alexeigurevich"
 
 # Setup a base system
 RUN apt-get update && \
     apt-get install -y curl wget g++ make libboost-all-dev git \
         tar gzip bzip2 build-essential \
-        python2.7-dev python-setuptools python-pip python-virtualenv \
+        python3-dev python3-setuptools python3-pip python3-virtualenv \
         zlib1g-dev default-jdk perl && \
     apt-get upgrade -y libstdc++6
 
 # Matplotlib dependencies
 RUN apt-get update && apt-get install -y pkg-config libfreetype6-dev \
-    libpng-dev python-matplotlib
+    libpng-dev python3-matplotlib
 
 # QUAST installation
 COPY . quast
-RUN pip install --upgrade setuptools pip && \
-    cd quast && \
-    python setup.py develop && \
+RUN cd quast && \
+    python3 setup.py develop && \
     cd ..

quast_libs/ca_utils/save_results.py

@@ -128,6 +128,8 @@ def save_result(result, report, fname, ref_fpath, genome_size):
         report.add_field(reporting.Fields.SUBSERROR, "%.2f" % (float(SNPs) * 100000.0 / float(aligned_assembly_bases)))
         report.add_field(reporting.Fields.INDELSERROR, "%.2f" % (float(report.get_field(reporting.Fields.INDELS))
                                                                  * 100000.0 / float(aligned_assembly_bases)))
+        report.add_field(reporting.Fields.ANI, "%.2f" % (100 - (float(SNPs) + float(report.get_field(reporting.Fields.INDELS)))
+                                                                / float(aligned_assembly_bases) * 100.0))
 
     # for misassemblies report:
     report.add_field(reporting.Fields.MIS_ALL_EXTENSIVE, region_misassemblies.count(Misassembly.RELOCATION) +

quast_libs/reporting.py

@@ -133,6 +133,7 @@ class Fields:
     MIS_LONG_INDELS = TAB + '# indels (> 5 bp)'
     UNCALLED = "# N's"
     UNCALLED_PERCENT = "# N's per 100 kbp"
+    ANI = "Average nucleotide identity (%)"
 
     # Genome statistics
     MAPPEDGENOME = 'Genome fraction (%)'
@@ -211,7 +212,7 @@ class Fields:
              MISLOCAL, MIS_SCAFFOLDS_GAP, MIS_LOCAL_SCAFFOLDS_GAP,
              STRUCT_VARIATIONS, POTENTIAL_MGE, UNALIGNED_MISASSEMBLED_CTGS,
              UNALIGNED, UNALIGNEDBASES, MAPPEDGENOME, DUPLICATION_RATIO, AVE_READ_SUPPORT,
-             UNCALLED_PERCENT, SUBSERROR, INDELSERROR, GENES, OPERONS,
+             UNCALLED_PERCENT, SUBSERROR, INDELSERROR, ANI, GENES, OPERONS,
              BUSCO_COMPLETE, BUSCO_PART,
              PREDICTED_GENES_UNIQUE, PREDICTED_GENES, RNA_GENES,
              LARGALIGN, TOTAL_ALIGNED_LEN, NA50, NGA50, NAx, NGAx, auNA, auNGA, LA50, LGA50, LAx, LGAx,
@@ -271,7 +272,7 @@ class Fields:
 
         ('Per base quality', [SUBSERROR, MISMATCHES,
                         INDELSERROR, INDELS, MIS_SHORT_INDELS, MIS_LONG_INDELS, INDELSBASES,
-                        UNCALLED_PERCENT, UNCALLED, ]),
+                        UNCALLED_PERCENT, UNCALLED, ANI, ]),
 
         ('Statistics without reference', [CONTIGS, CONTIGS__FOR_THRESHOLDS, LARGCONTIG, TOTALLEN, TOTALLENS__FOR_THRESHOLDS,
                                           N50, Nx, auN, L50, Lx, GC,]),
@@ -323,7 +324,7 @@ class Quality:
              BUSCO_COMPLETE,
              MAPPED_READS, MAPPED_READS_PCNT, PROPERLY_PAIRED_READS, PROPERLY_PAIRED_READS_PCNT,
              KMER_COMPLETENESS, KMER_CORR_LENGTH,
-             DEPTH, COVERAGE__FOR_THRESHOLDS],
+             DEPTH, COVERAGE__FOR_THRESHOLDS, ANI],
         Quality.LESS_IS_BETTER:
             [L50, LG50, Lx, LGx,
              MISLOCAL, MISASSEMBL, MISCONTIGS, MISCONTIGSBASES, MISINTERNALOVERLAP,