Comments (6)
Hi Tanveer,
Thank you for your interest in NanoCaller.
For HG001 (NA12878), we used aligned CCS reads from here: (https://ftp-trace.ncbi.nlm.nih.gov/giab/ftp/data/NA12878/PacBio_SequelII_CCS_11kb/HG001_GRCh38/) so we didn't do any alignment ourselves. For HG002-HG004, we obtained PacBio CCS FASTQ files from precisionFDA (now available at https://data.nist.gov/pdr/od/id/mds2-2336). We aligned these reads with minimap2 without any duplicate removal. Generally, for long-reads sequencing, you do not need to do any duplicate removal, regardless of PCR-free or not.
You can consult the supplementary file of precisionFDA challenge preprint to see what different teams did for alignment and any other pre-processing steps. Since we did not sequence these genomes, we do not if anything else was done before the FASTQ files were released publicly, so it would be better to consult with the sequencing team about that.
Just FYI, we used the following minimap2 command for CCS reads, with settings used in several aligned files provided by GIAB (example):
minimap2 -a -x map-pb -k 19 -O 5,56 -E 4,1 -B 5 -z 400,50 -r 2k --eqx --secondary=no
Let me know if there are any further questions.
from nanocaller.
Thank you so much Umair for your quick and detailed reply. It's clear now.
Would you like to mention a short flow of tools in ensemble mode (like mentioned here) for your precisionFDA challenge submission? I want to know if I can use PySaprk dataframes for SAM data to integrate such ensemble mode variant calling workflow. Thanks.
from nanocaller.
For our precisionFDA submission we had a somewhat elementary ensemble. We ran NanoCaller, Clair and Medaka separately on BAM files. After variant calling we took variant quality scores (QUAL field) of predictions made by each of the three variant callers and scaled them so that were all on a uniform scale. The reason for this is most of Clair's quality score were mostly in 500-1000, Medaka's score were mostly in 0-25, and NanoCaller's quality scores were mostly in 30-400 range.
We performed a weighted vote for each variant call, instead of majority voting. More precisely, we looked at each variant call in the union of three tools, issued the same call in final output with the following quality score:
QUAL=(NanoCaller QUAL score for the variant +Clair QUAL score for the variant +Medaka QUAL score for the variant)/3
If a tool did not make the variant call in union set we gave a quality score of 0 for that variant caller. So if there was a variant call made by NanoCaller with QUAL=240, but Medaka and Clair did not make this variant call, then the ensemble will issue the same variant call with QUAL=(214+0+0)/3=80.
So instead of removing calls with low number of votes, we just gave it a lower score. The reason for this is that there are several variant calls that are uniquely called by only one or two methods, so it would not seem prudent to discard such calls. This allows us maximum recall, but with quality score filtering we can make the callset more stringent to remove low vote supporting calls.
from nanocaller.
Thanks Umair for this detailed answer. Is it possible to contact you on your email for a possible collaboration? Thanks.
from nanocaller.
Hi Tanveer, yes that would be great. Could you give me your email so I can copy the rest of my team in the email as well?
from nanocaller.
Sure Umair. This is my email id: [email protected]
from nanocaller.
Related Issues (20)
- High coverage genotyping outputs only G HOT 10
- Using NanoCaller with BED files and Indel Calling Gives Empty VCF files HOT 3
- bcf tools error and error in whatshap HOT 4
- docker build fails due to missing NanoCaller_WGS file HOT 2
- Models compatible with Guppy6.x.x HOT 1
- Failed installation HOT 2
- bam file issue!
- what do you want to fill in the β--chromβ option? I read the instructions and there is no document. HOT 22
- header error in variant_calls.snps.phrased.vcf.gz HOT 6
- Unnormal results ! HOT 8
- AD (allelic depth) field HOT 5
- TypeError HOT 2
- Supporting reads HOT 3
- conda install incompatibilities HOT 2
- Models for R10.4.1 dorado
- Module not found tqtm after install with conda HOT 1
- Error phasing large chromosomes (550mb chromosome) due to inability to store as tbi index HOT 3
- Applications HOT 12
- indels HOT 2
- keras import issue by conda install
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
π Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. πππ
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google β€οΈ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from nanocaller.