def sample_status_note(project_name=None, flowcell=None, username=None, password=None, url=None, ordered_million_reads=None, uppnex_id=None, customer_reference=None, bc_count=None, project_alias=[], projectdb="projects", samplesdb="samples", flowcelldb="flowcells", phix=None, is_paired=True, **kw): """Make a sample status note. Used keywords: :param project_name: project name :param flowcell: flowcell id :param username: db username :param password: db password :param url: db url :param ordered_million_reads: number of ordered reads in millions :param uppnex_id: the uppnex id :param customer_reference: customer project name :param project_alias: project alias name :param phix: phix error rate :param is_paired: True if run is paired-end, False for single-end """ # Cutoffs cutoffs = { "phix_err_cutoff": 2.0, "qv_cutoff": 30, } instrument = _parse_instrument_config( os.path.expanduser(kw.get("instrument_config", ""))) instrument_dict = {i['instrument_id']: i for i in instrument} # parameters parameters = { "project_name": None, "start_date": None, "FC_id": None, "scilifelab_name": None, "rounded_read_count": None, "phix_error_rate": None, "avg_quality_score": None, "pct_q30_bases": None, "success": None, "run_mode": None, "is_paired": True } # key mapping from sample_run_metrics to parameter keys srm_to_parameter = { "project_name": "sample_prj", "FC_id": "flowcell", "scilifelab_name": "barcode_name", "start_date": "date", "rounded_read_count": "bc_count", "lane": "lane" } LOG.debug("got parameters {}".format(parameters)) output_data = { 'stdout': StringIO(), 'stderr': StringIO(), 'debug': StringIO() } if not _assert_flowcell_format(flowcell): LOG.warn( "Wrong flowcell format {}; skipping. Please use the flowcell id (format \"[A-Z0-9\-]+\")" .format(flowcell)) return output_data output_data = _update_sample_output_data(output_data, cutoffs) # Connect and run s_con = SampleRunMetricsConnection(dbname=samplesdb, username=username, password=password, url=url) fc_con = FlowcellRunMetricsConnection(dbname=flowcelldb, username=username, password=password, url=url) p_con = ProjectSummaryConnection(dbname=projectdb, username=username, password=password, url=url) # Set up paragraphs paragraphs = sample_note_paragraphs() headers = sample_note_headers() # Get project project = p_con.get_entry(project_name) source = p_con.get_info_source(project_name) if not project: LOG.warn("No such project '{}'".format(project_name)) return output_data # Set samples list sample_run_list = _set_sample_run_list(project_name, flowcell, project_alias, s_con) if len(sample_run_list) == 0: LOG.warn( "No samples for project '{}', flowcell '{}'. Maybe there are no sample run metrics in statusdb?" .format(project_name, flowcell)) return output_data # Set options ordered_million_reads = _literal_eval_option(ordered_million_reads) bc_count = _literal_eval_option(bc_count) phix = _literal_eval_option(phix) # Count number of times a sample has been run on a flowcell; if several, make lane-specific reports sample_count = Counter([x.get("barcode_name") for x in sample_run_list]) # Loop samples and collect information s_param_out = [] fcdoc = None for s in sample_run_list: s_param = {} LOG.debug( "working on sample '{}', sample run metrics name '{}', id '{}'". format(s.get("barcode_name", None), s.get("name", None), s.get("_id", None))) s_param.update(parameters) s_param.update( {key: s[srm_to_parameter[key]] for key in srm_to_parameter.keys()}) fc = "{}_{}".format(s.get("date"), s.get("flowcell")) # Get instrument try: s_param.update(instrument_dict[fc_con.get_instrument(str(fc))]) except: LOG.warn( "Failed to set instrument and software versions for flowcell {} in report due to missing RunInfo -> Instrument field in statusdb. Either rerun 'pm qc update-qc' or search-and-replace 'NN' in the sample report." .format(fc)) s_param.update(instrument_dict['default']) # Get run mode if not fcdoc or fcdoc.get("name") != fc: fcdoc = fc_con.get_entry(fc) runp = fcdoc.get("RunParameters", {}) s_param[ "sequencing_platform"] = "MiSeq" if "MCSVersion" in runp else "HiSeq2500" s_param["clustering_method"] = "onboard clustering" if runp.get( "ClusteringChoice", "") == "OnBoardClustering" or s_param[ "sequencing_platform"] == "MiSeq" else "cBot" s_param["sequencing_setup"] = fcdoc.get("run_setup") s_param["sequencing_mode"] = runp.get("RunMode", "High Output") s_param["sequencing_software"] = "RTA {}".format( runp.get("RTAVersion")) if s_param["sequencing_platform"] == "MiSeq": s_param["sequencing_software"] = "MCS {}/{}".format( runp.get("MCSVersion"), s_param["sequencing_software"]) else: s_param["sequencing_software"] = "{} {}/{}".format( runp.get("ApplicationName"), runp.get("ApplicationVersion"), s_param["sequencing_software"]) s_param["is_paired"] = fc_con.is_paired_end(str(fc)) if s_param["is_paired"] is None: LOG.warn( "Could not determine run setup for flowcell {}. Will assume paired-end." .format(fc)) s_param["is_paired"] = True s_param.update(software_versions) s_param["phix_error_rate"] = fc_con.get_phix_error_rate( str(fc), s["lane"]) if phix: s_param["phix_error_rate"] = _get_phix_error_rate(s["lane"], phix) # Get quality score from demultiplex stats, if that fails # (which it shouldn't), fall back on fastqc data. (avg_quality_score, pct_q30_bases) = fc_con.get_barcode_lane_statistics( project_name, s.get("barcode_name"), fc, s["lane"]) s_param[ 'avg_quality_score'] = avg_quality_score if avg_quality_score else calc_avg_qv( s) if not s_param['avg_quality_score']: LOG.warn( "Setting average quality failed for sample {}, id {}".format( s.get("name"), s.get("_id"))) s_param['pct_q30_bases'] = pct_q30_bases if not s_param['pct_q30_bases']: LOG.warn( "Setting % of >= Q30 Bases (PF) failed for sample {}, id {}". format(s.get("name"), s.get("_id"))) # Compare phix error and qv to cutoffs err_stat = "OK" qv_stat = "OK" if s_param["phix_error_rate"] > cutoffs["phix_err_cutoff"]: err_stat = "HIGH" elif s_param["phix_error_rate"] == -1: err_stat = "N/A" if s_param["avg_quality_score"] < cutoffs["qv_cutoff"]: qv_stat = "LOW" output_data["stdout"].write( "{:>18}\t{:>6}\t{:>12}\t{:>12}\t{:>12}\t{:>12}\n".format( s["barcode_name"], s["lane"], s_param["phix_error_rate"], err_stat, s_param["avg_quality_score"], qv_stat)) # Update/set remaning sample run parameters, falling back on project defaults if *key* is missing s_param['ordered_amount'] = s_param.get( 'ordered_amount', p_con.get_ordered_amount(project_name, samples=p_con.get_entry( project_name, 'samples'))) s_param['customer_reference'] = s_param.get( 'customer_reference', project.get('customer_reference')) s_param['uppnex_project_id'] = s_param.get('uppnex_project_id', project.get('uppnex_id')) # Override database settings if options passed at command line if ordered_million_reads: s_param["ordered_amount"] = _get_ordered_million_reads( s["barcode_name"], ordered_million_reads) if bc_count: s_param["rounded_read_count"] = _round_read_count_in_millions( _get_bc_count(s["barcode_name"], bc_count, s)) else: s_param["rounded_read_count"] = _round_read_count_in_millions( s_param["rounded_read_count"]) if uppnex_id: s_param["uppnex_project_id"] = uppnex_id if customer_reference: s_param["customer_reference"] = customer_reference # Get the project sample name corresponding to the sample run project_sample = p_con.get_project_sample( project_name, s.get("project_sample_name", None)) if project_sample: LOG.debug( "project sample run metrics mapping found: '{}' : '{}'".format( s["name"], project_sample["sample_name"])) project_sample_item = project_sample['project_sample'] # Set project_sample_d: a dictionary mapping from sample run metrics name to sample run metrics database id project_sample_d = _set_project_sample_dict( project_sample_item, source) if not project_sample_d: LOG.warn( "No sample_run_metrics information for sample '{}', barcode name '{}', id '{}'\n\tProject summary information {}" .format(s["name"], s["barcode_name"], s["_id"], project_sample)) # Check if sample run metrics name present in project database: if so, verify that database ids are consistent if s["name"] not in project_sample_d.keys(): LOG.warn( "no such sample run metrics '{}' in project sample run metrics dictionary" .format(s["name"])) else: if s["_id"] == project_sample_d[s["name"]]: LOG.debug( "project sample run metrics mapping found: '{}' : '{}'" .format(s["name"], project_sample_d[s["name"]])) else: LOG.warn( "inconsistent mapping for '{}': '{}' != '{}' (project summary id)" .format(s["name"], s["_id"], project_sample_d[s["name"]])) s_param['customer_name'] = project_sample_item.get( "customer_name", None) # Always normalize submitted id, since module textttable does not support unicode if type(s_param['customer_name']) is unicode: s_param['customer_name'] = unicodedata.normalize( 'NFKD', s_param['customer_name']).encode('ascii', 'ignore') # No project sample found. Manual upload to database necessary. else: s_param['customer_name'] = None LOG.warn( "No project sample name found for sample run name '{}'".format( s["barcode_name"])) LOG.info( "Please run 'pm qc upload-qc FLOWCELL_ID --extensive-matching' to update project sample names " ) LOG.info( "or 'pm qc update --sample_prj PROJECT_NAME --names BARCODE_TO_SAMPLE_MAP to update project sample names." ) LOG.info("Please refer to the pm documentation for examples.") query_ok(force=kw.get("force", False)) # Finally assess sequencing success, update parameters and set outputs s_param['success'] = sequencing_success(s_param, cutoffs) s_param.update({ k: "N/A" for k in s_param.keys() if s_param[k] is None or s_param[k] == "" or s_param[k] == -1.0 }) if sample_count[s.get("barcode_name")] > 1: outfile = "{}_{}_{}_{}.pdf".format(s["barcode_name"], s["date"], s["flowcell"], s["lane"]) else: outfile = "{}_{}_{}.pdf".format(s["barcode_name"], s["date"], s["flowcell"]) s_param["outfile"] = outfile s_param_out.append(s_param) # Write final output to reportlab and rst files output_data["debug"].write( json.dumps({ 's_param': s_param_out, 'sample_runs': {s["name"]: s["barcode_name"] for s in sample_run_list} })) notes = [ make_note(headers=headers, paragraphs=paragraphs, **sp) for sp in s_param_out ] rest_notes = make_sample_rest_notes( "{}_{}_{}_sample_summary.rst".format(project_name, s.get("date", None), s.get("flowcell", None)), s_param_out) concatenate_notes( notes, "{}_{}_{}_sample_summary.pdf".format(project_name, s.get("date", None), s.get("flowcell", None))) return output_data
def multiplex_qc(self): MAX_UNDEMULTIPLEXED_INDEX_COUNT = 1000000 EXPECTED_LANE_YIELD = 143000000 MAX_PHIX_ERROR_RATE = 2.0 MIN_PHIX_ERROR_RATE = 0.0 MIN_GTQ30 = 80.0 read_pairs = True out_data = [] if not self._check_pargs(['flowcell']): return url = self.pargs.url if self.pargs.url else self.app.config.get("db", "url") if not url: self.app.log.warn("Please provide a valid url: got {}".format(url)) return # Construct the short form of the fcid sp = os.path.basename(self.pargs.flowcell).split("_") fcid = "_".join([sp[0],sp[-1]]) # Get a connection to the flowcell database and fetch the corresponding document self.log.debug("Connecting to flowcell database".format(fcid)) fc_con = FlowcellRunMetricsConnection(dbname=self.app.config.get("db", "flowcells"), **vars(self.app.pargs)) self.log.debug("Fetching run metrics entry for flowcell {}".format(fcid)) fc_doc = fc_con.get_entry(fcid) if not fc_doc: self.log.warn("Could not fetch run metrics entry for flowcell {}".format(fcid)) return # Adjust the read pairs variable according to the run setup read_pairs = fc_con.is_paired_end(fcid) # Get the yield per sample from the Demultiplex_Stats self.log.debug("Getting yield for flowcell {}".format(fcid)) sample_yield = self._get_yield_per_sample(fc_doc, read_pairs) # Get the yield per lane from the Demultiplex_Stats self.log.debug("Getting lane yield for flowcell {}".format(fcid)) lane_yield = self._get_yield_per_lane(fc_doc, read_pairs) lanes = lane_yield.keys() # Get the number of samples in the pools from the Demultiplex_Stats self.log.debug("Getting lane pool sizes for flowcell {}".format(fcid)) pool_size = self._get_pool_size(fc_doc) # Get the sample information from the csv samplesheet self.log.debug("Getting csv samplesheet data for flowcell {}".format(fcid)) ssheet_samples = self._get_samplesheet_sample_data(fc_doc) if len(ssheet_samples) == 0: self.log.warn("No samplesheet data available for flowcell {}".format(fcid)) # Verify that all samples in samplesheet have reported metrics for id in ssheet_samples.keys(): for key in ssheet_samples[id].keys(): lane, index = key.split("_") project = ssheet_samples[id][key][0] if id not in sample_yield or \ key not in sample_yield[id]: self.log.warn("Sample {} from project {} is in samplesheet but no yield was reported in " \ "Demultiplex_Stats.htm for lane {} and index {}".format(id, project, lane, index)) continue sample_yield[id][key].append('verified') # Check that all samples in Demultiplex_Stats have entries in Samplesheet for id in sample_yield.keys(): for key in sample_yield[id].keys(): lane, index = key.split("_") if "verified" not in sample_yield[id][key] and \ index != "Undetermined": self.log.warn("Sample {} from project {}, with index {} on lane {} is in Demultiplex_Stats " \ "but no corresponding entry is present in SampleSheet".format(id, sample_yield[id][key][1], index, lane)) # Check the PhiX error rate for each lane self.log.debug("Getting PhiX error rates for flowcell {}".format(fcid)) for lane in lanes: status = "N/A" err_rate = fc_con.get_phix_error_rate(fcid,lane) if err_rate < 0: self.log.warn("Could not get PhiX error rate for lane {} on flowcell {}".format(lane,fcid)) elif err_rate <= MIN_PHIX_ERROR_RATE or err_rate > MAX_PHIX_ERROR_RATE: status = "FAIL" else: status = "PASS" out_data.append([status, "PhiX error rate", lane, err_rate, "{} < PhiX e (%) <= {}".format(MIN_PHIX_ERROR_RATE, MAX_PHIX_ERROR_RATE)]) # Check the %>=Q30 value for each sample sample_quality = self._get_quality_per_sample(fc_doc) for id in sample_quality.keys(): for key in sample_quality[id].keys(): lane, index = key.split("_") status = "FAIL" if float(sample_quality[id][key][0]) >= MIN_GTQ30: status = "PASS" out_data.append([status,"Sample quality",lane,sample_quality[id][key][2],id,sample_quality[id][key][0],"[%>=Q30 >= {}%]".format(MIN_GTQ30)]) # Check that each lane received the minimum amount of reads for lane, reads in lane_yield.items(): status = "FAIL" if reads >= EXPECTED_LANE_YIELD: status = "PASS" out_data.append([status,"Lane yield",lane,reads,"[Yield >= {}]".format(EXPECTED_LANE_YIELD)]) # Check that all samples in the pool have received a minimum number of reads for id in sample_yield.keys(): for key in sample_yield[id].keys(): lane, index = key.split("_") if index == "Undetermined": continue status = "FAIL" mplx_min = int(0.5*EXPECTED_LANE_YIELD/pool_size[lane]) if sample_yield[id][key][0] >= mplx_min: status = "PASS" out_data.append([status,"Sample yield",lane,sample_yield[id][key][1],id,sample_yield[id][key][0],"[Yield >= {}]".format(mplx_min)]) # Check that the number of undetermined reads in each lane is below 10% of the total yield for the lane for lane, reads in lane_yield.items(): status = "FAIL" key = "_".join([lane,"Undetermined"]) undetermined = sum([counts.get(key,[0])[0] for counts in sample_yield.values()]) cutoff = 0.1*reads if undetermined < cutoff: status = "PASS" out_data.append([status,"Index read",lane,undetermined,"[Undetermined < {}]".format(cutoff)]) # Check that no overrepresented index sequence exists in undemultiplexed output self.log.debug("Fetching undemultiplexed barcode data for flowcell {}".format(fcid)) undemux_data = self._get_undetermined_index_counts(fc_doc) if len(undemux_data) == 0: self.log.warn("No undemultiplexed barcode data available for flowcell {}".format(fcid)) for lane, counts in undemux_data.items(): mplx_min = int(min(MAX_UNDEMULTIPLEXED_INDEX_COUNT, 0.5*EXPECTED_LANE_YIELD/max(1,pool_size[lane]))) status = "N/A" if len(counts) > 0: for i in range(len(counts)): status = "FAIL" if int(counts[i][0]) < mplx_min: status = "PASS" out_data.append([status,"Index",lane,counts[i][1],counts[i][2],counts[i][0],"[Undetermined index < {}]".format(mplx_min)]) else: out_data.append([status,"Index",lane,"","",mplx_min,"-"]) self.app._output_data['stdout'].write("\n".join(["\t".join([str(r) for r in row]) for row in out_data]))
def sample_status_note(project_name=None, flowcell=None, username=None, password=None, url=None, ordered_million_reads=None, uppnex_id=None, customer_reference=None, bc_count=None, project_alias=[], projectdb="projects", samplesdb="samples", flowcelldb="flowcells", phix=None, is_paired=True, **kw): """Make a sample status note. Used keywords: :param project_name: project name :param flowcell: flowcell id :param username: db username :param password: db password :param url: db url :param ordered_million_reads: number of ordered reads in millions :param uppnex_id: the uppnex id :param customer_reference: customer project name :param project_alias: project alias name :param phix: phix error rate :param is_paired: True if run is paired-end, False for single-end """ # Cutoffs cutoffs = { "phix_err_cutoff" : 2.0, "qv_cutoff" : 30, } instrument = _parse_instrument_config(os.path.expanduser(kw.get("instrument_config",""))) instrument_dict = {i['instrument_id']: i for i in instrument} # parameters parameters = { "project_name" : None, "start_date" : None, "FC_id" : None, "scilifelab_name" : None, "rounded_read_count" : None, "phix_error_rate" : None, "avg_quality_score" : None, "pct_q30_bases" : None, "success" : None, "run_mode":None, "is_paired":True } # key mapping from sample_run_metrics to parameter keys srm_to_parameter = {"project_name":"sample_prj", "FC_id":"flowcell", "scilifelab_name":"barcode_name", "start_date":"date", "rounded_read_count":"bc_count", "lane": "lane"} LOG.debug("got parameters {}".format(parameters)) output_data = {'stdout':StringIO(), 'stderr':StringIO(), 'debug':StringIO()} if not _assert_flowcell_format(flowcell): LOG.warn("Wrong flowcell format {}; skipping. Please use the flowcell id (format \"[A-Z0-9\-]+\")".format(flowcell) ) return output_data output_data = _update_sample_output_data(output_data, cutoffs) # Connect and run s_con = SampleRunMetricsConnection(dbname=samplesdb, username=username, password=password, url=url) fc_con = FlowcellRunMetricsConnection(dbname=flowcelldb, username=username, password=password, url=url) p_con = ProjectSummaryConnection(dbname=projectdb, username=username, password=password, url=url) # Set up paragraphs paragraphs = sample_note_paragraphs() headers = sample_note_headers() # Get project project = p_con.get_entry(project_name) source = p_con.get_info_source(project_name) if not project: LOG.warn("No such project '{}'".format(project_name)) return output_data # Set samples list sample_run_list = _set_sample_run_list(project_name, flowcell, project_alias, s_con) if len(sample_run_list) == 0: LOG.warn("No samples for project '{}', flowcell '{}'. Maybe there are no sample run metrics in statusdb?".format(project_name, flowcell)) return output_data # Set options ordered_million_reads = _literal_eval_option(ordered_million_reads) bc_count = _literal_eval_option(bc_count) phix = _literal_eval_option(phix) # Count number of times a sample has been run on a flowcell; if several, make lane-specific reports sample_count = Counter([x.get("barcode_name") for x in sample_run_list]) # Loop samples and collect information s_param_out = [] fcdoc = None for s in sample_run_list: s_param = {} LOG.debug("working on sample '{}', sample run metrics name '{}', id '{}'".format(s.get("barcode_name", None), s.get("name", None), s.get("_id", None))) s_param.update(parameters) s_param.update({key:s[srm_to_parameter[key]] for key in srm_to_parameter.keys()}) fc = "{}_{}".format(s.get("date"), s.get("flowcell")) # Get instrument try: s_param.update(instrument_dict[fc_con.get_instrument(str(fc))]) except: LOG.warn("Failed to set instrument and software versions for flowcell {} in report due to missing RunInfo -> Instrument field in statusdb. Either rerun 'pm qc update-qc' or search-and-replace 'NN' in the sample report.".format(fc)) s_param.update(instrument_dict['default']) # Get run mode if not fcdoc or fcdoc.get("name") != fc: fcdoc = fc_con.get_entry(fc) runp = fcdoc.get("RunParameters",{}) s_param["sequencing_platform"] = "MiSeq" if "MCSVersion" in runp else "HiSeq2500" s_param["clustering_method"] = "onboard clustering" if runp.get("ClusteringChoice","") == "OnBoardClustering" or s_param["sequencing_platform"] == "MiSeq" else "cBot" s_param["sequencing_setup"] = fcdoc.get("run_setup") s_param["sequencing_mode"] = runp.get("RunMode","High Output") s_param["sequencing_software"] = "RTA {}".format(runp.get("RTAVersion")) if s_param["sequencing_platform"] == "MiSeq": s_param["sequencing_software"] = "MCS {}/{}".format(runp.get("MCSVersion"),s_param["sequencing_software"]) else: s_param["sequencing_software"] = "{} {}/{}".format(runp.get("ApplicationName"),runp.get("ApplicationVersion"),s_param["sequencing_software"]) s_param["is_paired"] = fc_con.is_paired_end(str(fc)) if s_param["is_paired"] is None: LOG.warn("Could not determine run setup for flowcell {}. Will assume paired-end.".format(fc)) s_param["is_paired"] = True s_param.update(software_versions) s_param["phix_error_rate"] = fc_con.get_phix_error_rate(str(fc), s["lane"]) if phix: s_param["phix_error_rate"] = _get_phix_error_rate(s["lane"], phix) # Get quality score from demultiplex stats, if that fails # (which it shouldn't), fall back on fastqc data. (avg_quality_score, pct_q30_bases) = fc_con.get_barcode_lane_statistics(project_name, s.get("barcode_name"), fc, s["lane"]) s_param['avg_quality_score'] = avg_quality_score if avg_quality_score else calc_avg_qv(s) if not s_param['avg_quality_score']: LOG.warn("Setting average quality failed for sample {}, id {}".format(s.get("name"), s.get("_id"))) s_param['pct_q30_bases'] = pct_q30_bases if not s_param['pct_q30_bases']: LOG.warn("Setting % of >= Q30 Bases (PF) failed for sample {}, id {}".format(s.get("name"), s.get("_id"))) # Compare phix error and qv to cutoffs err_stat = "OK" qv_stat = "OK" if s_param["phix_error_rate"] > cutoffs["phix_err_cutoff"]: err_stat = "HIGH" elif s_param["phix_error_rate"] == -1: err_stat = "N/A" if s_param["avg_quality_score"] < cutoffs["qv_cutoff"]: qv_stat = "LOW" output_data["stdout"].write("{:>18}\t{:>6}\t{:>12}\t{:>12}\t{:>12}\t{:>12}\n".format(s["barcode_name"], s["lane"], s_param["phix_error_rate"], err_stat, s_param["avg_quality_score"], qv_stat)) # Update/set remaning sample run parameters, falling back on project defaults if *key* is missing s_param['ordered_amount'] = s_param.get('ordered_amount', p_con.get_ordered_amount(project_name, samples=p_con.get_entry(project_name,'samples'))) s_param['customer_reference'] = s_param.get('customer_reference', project.get('customer_reference')) s_param['uppnex_project_id'] = s_param.get('uppnex_project_id', project.get('uppnex_id')) # Override database settings if options passed at command line if ordered_million_reads: s_param["ordered_amount"] = _get_ordered_million_reads(s["barcode_name"], ordered_million_reads) if bc_count: s_param["rounded_read_count"] = _round_read_count_in_millions(_get_bc_count(s["barcode_name"], bc_count, s)) else: s_param["rounded_read_count"] = _round_read_count_in_millions(s_param["rounded_read_count"]) if uppnex_id: s_param["uppnex_project_id"] = uppnex_id if customer_reference: s_param["customer_reference"] = customer_reference # Get the project sample name corresponding to the sample run project_sample = p_con.get_project_sample(project_name, s.get("project_sample_name", None)) if project_sample: LOG.debug("project sample run metrics mapping found: '{}' : '{}'".format(s["name"], project_sample["sample_name"])) project_sample_item = project_sample['project_sample'] # Set project_sample_d: a dictionary mapping from sample run metrics name to sample run metrics database id project_sample_d = _set_project_sample_dict(project_sample_item, source) if not project_sample_d: LOG.warn("No sample_run_metrics information for sample '{}', barcode name '{}', id '{}'\n\tProject summary information {}".format(s["name"], s["barcode_name"], s["_id"], project_sample)) # Check if sample run metrics name present in project database: if so, verify that database ids are consistent if s["name"] not in project_sample_d.keys(): LOG.warn("no such sample run metrics '{}' in project sample run metrics dictionary".format(s["name"]) ) else: if s["_id"] == project_sample_d[s["name"]]: LOG.debug("project sample run metrics mapping found: '{}' : '{}'".format(s["name"], project_sample_d[s["name"]])) else: LOG.warn("inconsistent mapping for '{}': '{}' != '{}' (project summary id)".format(s["name"], s["_id"], project_sample_d[s["name"]])) s_param['customer_name'] = project_sample_item.get("customer_name", None) # Always normalize submitted id, since module textttable does not support unicode if type(s_param['customer_name']) is unicode: s_param['customer_name'] = unicodedata.normalize('NFKD', s_param['customer_name']).encode('ascii', 'ignore') # No project sample found. Manual upload to database necessary. else: s_param['customer_name'] = None LOG.warn("No project sample name found for sample run name '{}'".format(s["barcode_name"])) LOG.info("Please run 'pm qc upload-qc FLOWCELL_ID --extensive-matching' to update project sample names ") LOG.info("or 'pm qc update --sample_prj PROJECT_NAME --names BARCODE_TO_SAMPLE_MAP to update project sample names.") LOG.info("Please refer to the pm documentation for examples.") query_ok(force=kw.get("force", False)) # Finally assess sequencing success, update parameters and set outputs s_param['success'] = sequencing_success(s_param, cutoffs) s_param.update({k:"N/A" for k in s_param.keys() if s_param[k] is None or s_param[k] == "" or s_param[k] == -1.0}) if sample_count[s.get("barcode_name")] > 1: outfile = "{}_{}_{}_{}.pdf".format(s["barcode_name"], s["date"], s["flowcell"], s["lane"]) else: outfile = "{}_{}_{}.pdf".format(s["barcode_name"], s["date"], s["flowcell"]) s_param["outfile"] = outfile s_param_out.append(s_param) # Write final output to reportlab and rst files output_data["debug"].write(json.dumps({'s_param': s_param_out, 'sample_runs':{s["name"]:s["barcode_name"] for s in sample_run_list}})) notes = [make_note(headers=headers, paragraphs=paragraphs, **sp) for sp in s_param_out] rest_notes = make_sample_rest_notes("{}_{}_{}_sample_summary.rst".format(project_name, s.get("date", None), s.get("flowcell", None)), s_param_out) concatenate_notes(notes, "{}_{}_{}_sample_summary.pdf".format(project_name, s.get("date", None), s.get("flowcell", None))) return output_data
def multiplex_qc(self): MAX_UNDEMULTIPLEXED_INDEX_COUNT = 1000000 EXPECTED_LANE_YIELD = 143000000 MAX_PHIX_ERROR_RATE = 2.0 MIN_PHIX_ERROR_RATE = 0.0 MIN_GTQ30 = 80.0 read_pairs = True out_data = [] if not self._check_pargs(['flowcell']): return url = self.pargs.url if self.pargs.url else self.app.config.get( "db", "url") if not url: self.app.log.warn("Please provide a valid url: got {}".format(url)) return # Construct the short form of the fcid sp = os.path.basename(self.pargs.flowcell).split("_") fcid = "_".join([sp[0], sp[-1]]) # Get a connection to the flowcell database and fetch the corresponding document self.log.debug("Connecting to flowcell database".format(fcid)) fc_con = FlowcellRunMetricsConnection(dbname=self.app.config.get( "db", "flowcells"), **vars(self.app.pargs)) self.log.debug( "Fetching run metrics entry for flowcell {}".format(fcid)) fc_doc = fc_con.get_entry(fcid) if not fc_doc: self.log.warn( "Could not fetch run metrics entry for flowcell {}".format( fcid)) return # Adjust the read pairs variable according to the run setup read_pairs = fc_con.is_paired_end(fcid) # Get the yield per sample from the Demultiplex_Stats self.log.debug("Getting yield for flowcell {}".format(fcid)) sample_yield = self._get_yield_per_sample(fc_doc, read_pairs) # Get the yield per lane from the Demultiplex_Stats self.log.debug("Getting lane yield for flowcell {}".format(fcid)) lane_yield = self._get_yield_per_lane(fc_doc, read_pairs) lanes = lane_yield.keys() # Get the number of samples in the pools from the Demultiplex_Stats self.log.debug("Getting lane pool sizes for flowcell {}".format(fcid)) pool_size = self._get_pool_size(fc_doc) # Get the sample information from the csv samplesheet self.log.debug( "Getting csv samplesheet data for flowcell {}".format(fcid)) ssheet_samples = self._get_samplesheet_sample_data(fc_doc) if len(ssheet_samples) == 0: self.log.warn( "No samplesheet data available for flowcell {}".format(fcid)) # Verify that all samples in samplesheet have reported metrics for id in ssheet_samples.keys(): for key in ssheet_samples[id].keys(): lane, index = key.split("_") project = ssheet_samples[id][key][0] if id not in sample_yield or \ key not in sample_yield[id]: self.log.warn("Sample {} from project {} is in samplesheet but no yield was reported in " \ "Demultiplex_Stats.htm for lane {} and index {}".format(id, project, lane, index)) continue sample_yield[id][key].append('verified') # Check that all samples in Demultiplex_Stats have entries in Samplesheet for id in sample_yield.keys(): for key in sample_yield[id].keys(): lane, index = key.split("_") if "verified" not in sample_yield[id][key] and \ index != "Undetermined": self.log.warn("Sample {} from project {}, with index {} on lane {} is in Demultiplex_Stats " \ "but no corresponding entry is present in SampleSheet".format(id, sample_yield[id][key][1], index, lane)) # Check the PhiX error rate for each lane self.log.debug("Getting PhiX error rates for flowcell {}".format(fcid)) for lane in lanes: status = "N/A" err_rate = fc_con.get_phix_error_rate(fcid, lane) if err_rate < 0: self.log.warn( "Could not get PhiX error rate for lane {} on flowcell {}". format(lane, fcid)) elif err_rate <= MIN_PHIX_ERROR_RATE or err_rate > MAX_PHIX_ERROR_RATE: status = "FAIL" else: status = "PASS" out_data.append([ status, "PhiX error rate", lane, err_rate, "{} < PhiX e (%) <= {}".format(MIN_PHIX_ERROR_RATE, MAX_PHIX_ERROR_RATE) ]) # Check the %>=Q30 value for each sample sample_quality = self._get_quality_per_sample(fc_doc) for id in sample_quality.keys(): for key in sample_quality[id].keys(): lane, index = key.split("_") status = "FAIL" if float(sample_quality[id][key][0]) >= MIN_GTQ30: status = "PASS" out_data.append([ status, "Sample quality", lane, sample_quality[id][key][2], id, sample_quality[id][key][0], "[%>=Q30 >= {}%]".format(MIN_GTQ30) ]) # Check that each lane received the minimum amount of reads for lane, reads in lane_yield.items(): status = "FAIL" if reads >= EXPECTED_LANE_YIELD: status = "PASS" out_data.append([ status, "Lane yield", lane, reads, "[Yield >= {}]".format(EXPECTED_LANE_YIELD) ]) # Check that all samples in the pool have received a minimum number of reads for id in sample_yield.keys(): for key in sample_yield[id].keys(): lane, index = key.split("_") if index == "Undetermined": continue status = "FAIL" mplx_min = int(0.5 * EXPECTED_LANE_YIELD / pool_size[lane]) if sample_yield[id][key][0] >= mplx_min: status = "PASS" out_data.append([ status, "Sample yield", lane, sample_yield[id][key][1], id, sample_yield[id][key][0], "[Yield >= {}]".format(mplx_min) ]) # Check that the number of undetermined reads in each lane is below 10% of the total yield for the lane for lane, reads in lane_yield.items(): status = "FAIL" key = "_".join([lane, "Undetermined"]) undetermined = sum( [counts.get(key, [0])[0] for counts in sample_yield.values()]) cutoff = 0.1 * reads if undetermined < cutoff: status = "PASS" out_data.append([ status, "Index read", lane, undetermined, "[Undetermined < {}]".format(cutoff) ]) # Check that no overrepresented index sequence exists in undemultiplexed output self.log.debug( "Fetching undemultiplexed barcode data for flowcell {}".format( fcid)) undemux_data = self._get_undetermined_index_counts(fc_doc) if len(undemux_data) == 0: self.log.warn( "No undemultiplexed barcode data available for flowcell {}". format(fcid)) for lane, counts in undemux_data.items(): mplx_min = int( min(MAX_UNDEMULTIPLEXED_INDEX_COUNT, 0.5 * EXPECTED_LANE_YIELD / max(1, pool_size[lane]))) status = "N/A" if len(counts) > 0: for i in range(len(counts)): status = "FAIL" if int(counts[i][0]) < mplx_min: status = "PASS" out_data.append([ status, "Index", lane, counts[i][1], counts[i][2], counts[i][0], "[Undetermined index < {}]".format(mplx_min) ]) else: out_data.append([status, "Index", lane, "", "", mplx_min, "-"]) self.app._output_data['stdout'].write("\n".join( ["\t".join([str(r) for r in row]) for row in out_data]))