vidjil issueshttps://gitlab.inria.fr/vidjil/vidjil/-/issues2024-01-19T19:03:58+01:00https://gitlab.inria.fr/vidjil/vidjil/-/issues/5145Upload by network; merge create hard file and no symlink2024-01-19T19:03:58+01:00THONIER FlorianUpload by network; merge create hard file and no symlinkIf you use a network to load data and choose a preprocess with merge, the resulting file will be store as a real file and no as a symlink. In this case, it will take more space.If you use a network to load data and choose a preprocess with merge, the resulting file will be store as a real file and no as a symlink. In this case, it will take more space.Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5144Large amount of memory used by Vidjil-algo2023-05-12T16:09:40+02:00Mikaël SalsonLarge amount of memory used by Vidjil-algoUn étudiant me signale que sur une machine avec 30G de RAM, Vidjil-algo utilise trop de mémoire : DRR346909, DRR346911 et DRR346910. Les jeux de données font une trentaine de Go.Un étudiant me signale que sur une machine avec 30G de RAM, Vidjil-algo utilise trop de mémoire : DRR346909, DRR346911 et DRR346910. Les jeux de données font une trentaine de Go.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5143Add D primers in the set of primers2023-05-11T17:49:33+02:00Mikaël SalsonAdd D primers in the set of primersAs requested by ~"NAN - Nantes" during 2023 Vidjil day.
Beware to the prioritization of primers (V/J primers should have the priority over D).As requested by ~"NAN - Nantes" during 2023 Vidjil day.
Beware to the prioritization of primers (V/J primers should have the priority over D).Web 2023.10THONIER FlorianTHONIER Florianhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5142Huge lengths in the read length distribution2023-06-21T15:20:21+02:00Mikaël SalsonHuge lengths in the read length distributionIn those samples, the read length distribution plot ranges from 0 to 6,500, which seems quite a lot: https://app.vidjil.org/57055-2?plot=Reads%20length,J/3%27%20gene,bar
The data doesn't look like 3rd generation sequencing.In those samples, the read length distribution plot ranges from 0 to 6,500, which seems quite a lot: https://app.vidjil.org/57055-2?plot=Reads%20length,J/3%27%20gene,bar
The data doesn't look like 3rd generation sequencing.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5141Read length distribution shows wrong sizes2024-01-22T15:22:21+01:00Mikaël SalsonRead length distribution shows wrong sizesOn these samples, I find that the read length distribution plot starts at -10% (!): https://app.vidjil.org/57056-25?plot=Reads%20length,Size,bar
It seems that the labels are not at their correct place. Even the main clone which is at 36...On these samples, I find that the read length distribution plot starts at -10% (!): https://app.vidjil.org/57056-25?plot=Reads%20length,Size,bar
It seems that the labels are not at their correct place. Even the main clone which is at 36%, only reaches 20% on the plot.Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5138Upload: Avertir dans le client quand les noms de fichiers ne respectent pas R...2023-05-04T15:44:06+02:00Mathieu GiraudUpload: Avertir dans le client quand les noms de fichiers ne respectent pas R1/R2
Normalement, on s'attend à exactement le même nom, avec juste R1 qui est changé en R2.
Si ce n'est pas le cas, avertir.
Soucis humains possibles qui seront diagnostiqués ainsi:
- deux fois le même fichier R1 (et ensuite flash fait des ...
Normalement, on s'attend à exactement le même nom, avec juste R1 qui est changé en R2.
Si ce n'est pas le cas, avertir.
Soucis humains possibles qui seront diagnostiqués ainsi:
- deux fois le même fichier R1 (et ensuite flash fait des choses, mais on ne détecte pas d'où vient le problème)
- ou on se plante de fichier R1/R2
- ou on inverse R1/R2https://gitlab.inria.fr/vidjil/vidjil/-/issues/5136Warnings: documenter pour les usagers2023-05-03T14:40:26+02:00Mathieu GiraudWarnings: documenter pour les usagersOn a https://www.vidjil.org/doc/warnings/ (générée à partir de warning.md), mais c'est encore un peu cryptique
Avoir une explication orientée usager, 1-3 phrase par warning, surtout pour ceux qui sont "fréquents". D'ailleurs, lesquels s...On a https://www.vidjil.org/doc/warnings/ (générée à partir de warning.md), mais c'est encore un peu cryptique
Avoir une explication orientée usager, 1-3 phrase par warning, surtout pour ceux qui sont "fréquents". D'ailleurs, lesquels sont les plus fréquents ?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5134Refresh admin.md documentation2024-01-22T15:27:45+01:00Mikaël SalsonRefresh admin.md documentation@flothoni points some inconsistencies in the `admin.md` documentation.@flothoni points some inconsistencies in the `admin.md` documentation.Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5133No log in database2024-01-22T15:28:47+01:00Mikaël SalsonNo log in databaseIt appears that we are not logging actions anymore in the database, since 52158ac0 (which came a few days after this attempt: 6fce80e1db).
Anyone knows why?
It should be fixed.It appears that we are not logging actions anymore in the database, since 52158ac0 (which came a few days after this attempt: 6fce80e1db).
Anyone knows why?
It should be fixed.Web 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5131Algo: Limit or warn clustering of sequences of too different lengths2023-04-03T12:34:58+02:00Mathieu GiraudAlgo: Limit or warn clustering of sequences of too different lengths
Very interesting cases reported by @Anne.
Some clones merged sequences with very different lengths, two peaks (biological ? artefact ?).
"no-clonality" already allows to check that, but we need some mechanism to prevent that or at leas...
Very interesting cases reported by @Anne.
Some clones merged sequences with very different lengths, two peaks (biological ? artefact ?).
"no-clonality" already allows to check that, but we need some mechanism to prevent that or at least to warn the userhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5130Améliorer la doc sur CloneDB2023-04-03T11:02:21+02:00Mathieu GiraudAméliorer la doc sur CloneDBCe n'est plus expérimental, cela fonctionne.
Un screenshot.
Une ou deux phrases pour usage bio (surtout primers).
https://www.vidjil.org/doc/user/#detailed-information-from-clonedbCe n'est plus expérimental, cela fonctionne.
Un screenshot.
Une ou deux phrases pour usage bio (surtout primers).
https://www.vidjil.org/doc/user/#detailed-information-from-clonedbhttps://gitlab.inria.fr/vidjil/vidjil/-/issues/5129Provide information when marked_pos (CDR3) is not found2024-02-01T06:04:02+01:00Mikaël SalsonProvide information when marked_pos (CDR3) is not foundIt may happen that we don't find the CDR3 because there are too many deletions either in the V or the J genes that go beyond the marked_pos.
In such a case the CDR3 will not be found and no reason will be provided for that.
More generall...It may happen that we don't find the CDR3 because there are too many deletions either in the V or the J genes that go beyond the marked_pos.
In such a case the CDR3 will not be found and no reason will be provided for that.
More generally if the marked_pos is not set in the germline, the user doesn't have that information. The user can only see that no CDR3 was found with no reason given for that (which is not the case for productivity).
We should give some information about it.
Here is an example where the marked_pos is deleted (appears in position 277/290 of the V gene but we have 16 deletions in the V).
```
GCTGTCATCTCTCAAAAGCCAAGCAGGGATATCTGTCAACGTGGAACCTCCCTGACGATCCAGTGTCAAGTCGATAGCCAAGTCACCATGATGTTCTGGTACCGTCAGCAACCTGGACAGAGCCTGACACTGATCGCAACTGCAAATCAGGGCTCTGAGGCCACATATGAGAGTGGATTTGTCATTGACAAGTTTCCCATCAGCCGCCCAAACCTAACATTCTCAACTCTGACTGTGAGCAACATGAGCCCTGAAGACAGCAGCATAAATCCCCTTGGGGGTCCCTATAATTCACCCCTCCACTTTGGGAACGGGACCAGGCTCACTGTGACAGGTATGGGGGCTCCACTCTTGACTCGGGGGTGCCTGGGTTTGACTG
```Algo 2024.04https://gitlab.inria.fr/vidjil/vidjil/-/issues/5127docker-compose, use env and more config for mysql2023-03-14T10:40:26+01:00THONIER Floriandocker-compose, use env and more config for mysqlMysql as used for the moment is not optimal as we need to change right on volume directory to used it (and it is not documented).
A realize that we can set more parameter in docker-compose to use it directly.
We can also use .env file...Mysql as used for the moment is not optimal as we need to change right on volume directory to used it (and it is not documented).
A realize that we can set more parameter in docker-compose to use it directly.
We can also use .env file to store some variable easily.https://gitlab.inria.fr/vidjil/vidjil/-/issues/5124Fuse; How to merge clonotype with various windows/ids length ?2023-03-01T10:54:49+01:00THONIER FlorianFuse; How to merge clonotype with various windows/ids length ?I was looking on analysis made with multiple configurations on the same sample file.
Window length parameter is not constant and the fuse don't work.
How can we resolve this ? I was naively thinking about `if the smaller window fit ins...I was looking on analysis made with multiple configurations on the same sample file.
Window length parameter is not constant and the fuse don't work.
How can we resolve this ? I was naively thinking about `if the smaller window fit inside the bigger we can merge`. But in this case, with the longer windows configuration, a clonotype on configuration `multi` (windows of 60nt length) is splitted into multiple smaller clonotype for configuration `clonality` (full sequence windows).
It is a one-to-many relationship. If it is easy to create a link at the fuse step, how to use it in client ?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5123Warning view; Ugly display on chrome2023-03-28T16:14:38+02:00THONIER FlorianWarning view; Ugly display on chromeWhile I was testing to take screenshot for documentation, I found that warning view is not similar that the view on firefox.
![Screenshot_20230228_102853](/uploads/7e5eff9708e63deadfb4961ff384b425/Screenshot_20230228_102853.png)While I was testing to take screenshot for documentation, I found that warning view is not similar that the view on firefox.
![Screenshot_20230228_102853](/uploads/7e5eff9708e63deadfb4961ff384b425/Screenshot_20230228_102853.png)Web 2023.10https://gitlab.inria.fr/vidjil/vidjil/-/issues/5122Problème analyse no-clustering2023-03-01T14:49:53+01:00Anne de SeptenvilleProblème analyse no-clusteringBonjour,
Je suis en train d'analyser pas mal de cDNA avec l'analyse no-clustering et je voulais vous signaler un sample bizarre pour lequel j'ai 94% des reads analysés en IGH et seulement 3% avec l'analyse no-clustering. Pour tout mes ...Bonjour,
Je suis en train d'analyser pas mal de cDNA avec l'analyse no-clustering et je voulais vous signaler un sample bizarre pour lequel j'ai 94% des reads analysés en IGH et seulement 3% avec l'analyse no-clustering. Pour tout mes autres patients les % de reads analysés sont sensiblement identiques.
Analyse IGH
https://app.vidjil.org/53787-2?
Analyse no-clustering
https://app.vidjil.org/53787-50?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5121Mise à jour des germlines sur serveur de prod2023-02-08T14:12:39+01:00Mikaël SalsonMise à jour des germlines sur serveur de prodDans #5120 je m'apperçois que les germlines ne sont pas à jour. Ils sont pris depuis le dossier du repo de docker et ils ne sont pas pris dans le dossier de vidjil-algo.
Autant ceux de vidjil-algo sont à jour, autant il n'y a pas de rai...Dans #5120 je m'apperçois que les germlines ne sont pas à jour. Ils sont pris depuis le dossier du repo de docker et ils ne sont pas pris dans le dossier de vidjil-algo.
Autant ceux de vidjil-algo sont à jour, autant il n'y a pas de raison pour que ceux du repo de docker soient à jour (ou en tout cas il faudrait faire un `make germline`) pour qu'ils le soient.
Utiliser les germlines du vidjil-algo a l'avantage de les avoir à jour mais l'inconvénient c'est qu'on peut moins facilement mettre à disposition des germlines demandées par les utilisateurs (situation qu'on a rencontrée).https://gitlab.inria.fr/vidjil/vidjil/-/issues/5120Erreur d'identification du V2023-02-08T13:07:16+01:00Anne de SeptenvilleErreur d'identification du VDans le cas de ce patient : https://app.vidjil.org/55797-2?
Pour le clone non productif, Vidjil donne un V3-7 alors que IMGT et IgBlast identifient tous les deux un V3-41 à 100%
Quand j'aligne la séquence avec celle du V3-7 je trouve...Dans le cas de ce patient : https://app.vidjil.org/55797-2?
Pour le clone non productif, Vidjil donne un V3-7 alors que IMGT et IgBlast identifient tous les deux un V3-41 à 100%
Quand j'aligne la séquence avec celle du V3-7 je trouve plein de différences...https://gitlab.inria.fr/vidjil/vidjil/-/issues/5119How to store value inside warnings ?2023-01-31T14:37:03+01:00THONIER FlorianHow to store value inside warnings ?Some warnings include more precise text than just warning description. For example `Bad evalue; 0.2`or something of this type.
On an external script I made, some warnings are throw with this type of behavior. Should we ad a field in wa...Some warnings include more precise text than just warning description. For example `Bad evalue; 0.2`or something of this type.
On an external script I made, some warnings are throw with this type of behavior. Should we ad a field in warning for this?https://gitlab.inria.fr/vidjil/vidjil/-/issues/5117artifact sequence; be able to substract them from number of reads2023-01-25T17:30:48+01:00THONIER Florianartifact sequence; be able to substract them from number of readsWhen a clone is tagged as noise/artifact, we should offer the possiblity to subsrat is value from what is proportion of reads shown.
ex: IGH D7-27//J1. If it is seen at 50% and hidden or tagged as noise, other clone still at the same s...When a clone is tagged as noise/artifact, we should offer the possiblity to subsrat is value from what is proportion of reads shown.
ex: IGH D7-27//J1. If it is seen at 50% and hidden or tagged as noise, other clone still at the same size.