Merge pull request #42 from nleroy917/seeding

Set Seed + Iterations

.vscode/settings.json

@@ -1,2 +1,11 @@
 {
-}
+    "css.validate": false,
+    "editor.tabSize": 2,
+    "editor.codeActionsOnSave": {
+      "source.fixAll": true,
+      "source.fixAll.eslint": true
+    },
+    "editor.formatOnSave": true, // Tell VSCode to format files on save
+    "editor.defaultFormatter": "esbenp.prettier-vscode" // Tell VSCode to use Prettier as default file formatter
+  }
+

optipyzer/const.py

@@ -1,4 +1,4 @@
-VERSION = "0.1.4"
+VERSION = "0.2.0"
 
 VALID_AMINO_ACIDS = "ARNDCQEGHILKMFPSTWYV"
 

optipyzer/oengine.py

@@ -428,7 +428,7 @@ def validate_query(query, DNA):
     return query, stop_pos
 
 
-def optimize_sequence(random_num_table, query):
+def optimize_sequence(random_num_table, query, seed=None):
     """
     takes the current iteration of the multi-species optimized codon preference table and uses it with a weighted
     codon-randomization method to convert a fasta-formatted protein sequence to an optimized DNA sequence
@@ -439,11 +439,17 @@ def optimize_sequence(random_num_table, query):
     :param query: a fasta=formatted protein sequence of the gene to be optimized
     :return:
     """
+    # set a seed if passed
+    # to create a random number generator
+    if seed:
+        rng = random.Random(seed)
+    else:
+        rng = random.Random()
     # initialize the DNA sequence
     optimmized_query = ""
     # loops through the query and generates a random integer between 1 and 100000 for each residue in query
     for residue in query:
-        value = random.randint(1, 100000001)
+        value = rng.randint(1, 100000001)
         # compares random number to the random number bounds for that codon
         # adds the appropriate DNA codon (based on the random number) to the optimized query
         for codon in random_num_table[residue]:
@@ -455,7 +461,7 @@ def optimize_sequence(random_num_table, query):
                 optimmized_query += codon
     # generates a random integer between 1 and 100000 and compares it to the random number bounds for the stop codons
     # adds the selected stop codon to the optimized DNA sequence
-    value = random.randint(1, 100000001)
+    value = rng.randint(1, 100000001)
     for codon in random_num_table["Stop"]:
         if (
             random_num_table["Stop"][codon][0]
@@ -660,7 +666,13 @@ def get_redundantaa_rn(query):
 
 
 def adjust_table(
-    rca_expression_dif, species_expression, et, aa_rn, query_table, multi_table
+    rca_expression_dif,
+    species_expression,
+    et,
+    aa_rn,
+    query_table,
+    multi_table,
+    seed=None,
 ):
     """
     Adjusts the table in favor of or against species that have a predicted expression different than their target
@@ -685,6 +697,10 @@ def adjust_table(
     encode that residue are the key, and the value is the codon preference after adjusting for species over- or under-
     performing.
     """
+    if seed is not None:
+        rng = random.Random(seed)
+    else:
+        rng = random.Random()
     for species in rca_expression_dif:
         # when current table is performing worse than the current best table, adjusts the multi_table
         # codon preferences in favor of species which currently has an expression difference greater than
@@ -696,7 +712,7 @@ def adjust_table(
             # in the query
             while aa_adjusted < 10:
                 aa_adjusted += 1
-                v = random.randint(1, 100000001)
+                v = rng.randint(1, 100000001)
                 for residue in aa_rn:
                     new_sum = 0
                     if aa_rn[residue][0] <= v < aa_rn[residue][1]:
@@ -722,6 +738,7 @@ def optimize_multitable_sd(
     species_expression,
     et=0.05,
     iterations=1000,
+    seed=None,
 ):
     """
     iterates upon the multi_table while optimizing the query to select the best-optimized DNA sequence using a sum of
@@ -756,7 +773,7 @@ def optimize_multitable_sd(
         # levels
         square_diff = 0
         # calls to optimmize the query sequence
-        optimized_seq = optimize_sequence(rn, query)
+        optimized_seq = optimize_sequence(rn, query, seed=seed)
         # calculates the rca measure of relative expression for each species
         rca = calculate_predicted_expression(rca_xyz, optimized_seq)
         # initializes a dictionary to store the difference in species expression
@@ -781,6 +798,7 @@ def optimize_multitable_sd(
                 aa_rn,
                 query_table,
                 multi_table,
+                seed=seed,
             )
             # gets a new random number table for the new table
             rn = get_multitable_randomnumbers(multi_table)
@@ -801,6 +819,7 @@ def optimize_multitable_ad(
     species_expression,
     et=0.05,
     iterations=1000,
+    seed=None,
 ):
     """
     iterates upon the multi_table while optimizing the query to select the best-optimized DNA sequence using an
@@ -860,6 +879,7 @@ def optimize_multitable_ad(
                 aa_rn,
                 query_table,
                 multi_table,
+                seed=seed,
             )
             # gets a new random number table for the new table
             rn = get_multitable_randomnumbers(multi_table)

optipyzer/optimization.py

@@ -1,5 +1,6 @@
+from typing import Union
 from optipyzer.const import DEFAULT_NUM_ITERATIONS
-from .oengine import (
+from optipyzer.oengine import (
     averaged_table,
     remove_prohibited_codons,
     find_prohibited_codons,
@@ -8,8 +9,8 @@
     optimize_multitable_ad,
     optimize_multitable_sd,
 )
-from .utils import seq_detect, aa_to_dna
-from .db.interfaces import calc_codon_usage
+from optipyzer.utils import seq_detect, aa_to_dna
+from optipyzer.db.interfaces import calc_codon_usage
 
 
 def _calc_average_table(usage_data: dict, weights: dict):
@@ -28,6 +29,7 @@ def codon_optimize(
     weights: dict = None,
     seq_type: str = None,
     iterations: int = DEFAULT_NUM_ITERATIONS,
+    seed: Union[str, int] = None,
 ):
     """Optimize a sequence given an organism list and a map/dictionary of weights"""
     if seq_type is None:
@@ -51,16 +53,34 @@ def codon_optimize(
 
     peptide_seq, stop_codon = validate_query(seq, (seq_type == "dna"))
 
+    # squared difference optimization
     (
         optimized_sd,
         min_difference_sumsquares,
         best_expression_sd,
     ) = optimize_multitable_sd(
-        average_table, peptide_seq, usage_data, rca_xyz, weights, iterations=iterations
+        average_table,
+        peptide_seq,
+        usage_data,
+        rca_xyz,
+        weights,
+        iterations=iterations,
+        seed=seed,
     )
 
-    optimized_ad, min_difference_absvalue, best_expression_ad = optimize_multitable_ad(
-        average_table, peptide_seq, usage_data, rca_xyz, weights, iterations=iterations
+    # absolute difference optimization
+    (
+        optimized_ad,
+        min_difference_absvalue,
+        best_expression_ad,
+    ) = optimize_multitable_ad(
+        average_table,
+        peptide_seq,
+        usage_data,
+        rca_xyz,
+        weights,
+        iterations=iterations,
+        seed=seed,
     )
 
     return {

optipyzer/request_models.py

@@ -1,4 +1,4 @@
-from typing import Dict, Optional
+from typing import Dict, Optional, Union
 from fastapi import Query
 from pydantic import BaseModel
 
@@ -7,6 +7,7 @@ class OptimizeQuery(BaseModel):
     seq: str
     weights: Dict[str, int]
     iterations: Optional[int]
+    seed: Optional[Union[int, str]]
 
     class Config:
         schema_extra = {

optipyzer/routers/optimize.py

@@ -27,6 +27,7 @@ def optimize_dna(query: OptimizeQuery = Depends(verify_dna)):
         query.weights,
         seq_type="dna",
         iterations=(query.iterations or DEFAULT_NUM_ITERATIONS),
+        seed=query.seed,
     )
 
     return result
@@ -49,6 +50,7 @@ def optimize_protein(query: OptimizeQuery = Depends(verify_protein)):
         query.weights,
         seq_type="protein",
         iterations=(query.iterations or DEFAULT_NUM_ITERATIONS),
+        seed=query.seed,
     )
 
     return result

sdk/optipyzer/api.py

@@ -1,15 +1,20 @@
-from typing import Dict
+from typing import Dict, Optional, Union
 import requests
 import time
-from .const import LOCAL_SERVER_BASE, PUBLIC_SERVER_BASE, SESSION_HDRS, SLEEP_MIN
-from .log import _LOGGER
-from .helpers import verify_dna, verify_protein
+from sdk.optipyzer.const import (
+    LOCAL_SERVER_BASE,
+    PUBLIC_SERVER_BASE,
+    SESSION_HDRS,
+    SLEEP_MIN,
+)
+from sdk.optipyzer.log import _LOGGER
+from sdk.optipyzer.helpers import verify_dna, verify_protein
 
 # return types
 from requests import Response
-from .models import SearchResult
-from .const import VALID_SEQ_TYPES
-from .models import CodonUsage, OptimizationResult
+from sdk.optipyzer.models import SearchResult
+from sdk.optipyzer.const import VALID_SEQ_TYPES
+from sdk.optipyzer.models import CodonUsage, OptimizationResult
 
 
 class api:
@@ -75,7 +80,12 @@ def search(self, name: str, limit: int = 50) -> SearchResult:
         return search_results
 
     def optimize(
-        self, seq: str, weights: Dict[str, int], seq_type: str = "dna"
+        self,
+        seq: str,
+        weights: Dict[str, int],
+        seq_type: str = "dna",
+        iterations: Optional[int] = None,
+        seed: Optional[Union[int, str]] = None,
     ) -> OptimizationResult:
         """Optimize a sequence given specific organism weights"""
         # force seq_type lower
@@ -93,7 +103,12 @@ def optimize(
         result = self._make_request(
             f"/optimize/{seq_type}",
             method="POST",
-            body_={"seq": seq, "weights": weights},
+            body_={
+                "seq": seq,
+                "weights": weights,
+                "iterations": iterations,
+                "seed": seed,
+            },
         )
         return result.json()
 

web/src/components/HyperParameterSelection.tsx

@@ -0,0 +1,38 @@
+import { Dispatch, FC, SetStateAction } from 'react'
+
+interface Props {
+  iterations: number | null
+  setIterations: Dispatch<SetStateAction<number | null>>
+  seed: number | string | null
+  setSeed: Dispatch<SetStateAction<number | string | null>>
+}
+
+const HyperParameterSelection: FC<Props> = (props) => {
+  const { iterations, setIterations, seed, setSeed } = props
+  return (
+    <div className="flex flex-col md:flex-row">
+      <div>
+        <p className="mt-2 text-lg font-bold text-blue-600">
+          Number of iterations:
+        </p>
+        <input
+          value={iterations || undefined}
+          onChange={(e) => setIterations(parseInt(e.target.value))}
+          placeholder="1000"
+          type="number"
+          className="p-2 mb-2 text-sm text-gray-600 border border-black rounded-lg shadow-md"
+        />
+      </div>
+      <div className="md:ml-2">
+        <p className="mt-2 text-lg font-bold text-blue-600">Random Seed:</p>
+        <input
+          value={seed || undefined}
+          onChange={(e) => setSeed(e.target.value)}
+          className="p-2 mb-2 text-sm text-gray-600 border border-black rounded-lg shadow-md"
+        />
+      </div>
+    </div>
+  )
+}
+
+export default HyperParameterSelection

web/src/pages/optimize.tsx

@@ -13,13 +13,16 @@ import WeightSelector from '@/components/WeightSelector'
 import { QueryError, QueryResult, Species, Weights } from '@/..'
 import { purgeWeights } from '@/utils/purgeWeights'
 import { NextPage } from 'next'
+import HyperParameterSelection from '@/components/HyperParameterSelection'
 
 const OptimizePage: NextPage = () => {
   // state
   const [seqType, setSeqType] = useState<string>('dna')
   const [species, setSpecies] = useState<Species[]>([])
   const [weights, setWeights] = useState<Weights>({})
   const [seq, setSeq] = useState<string>('')
+  const [iterations, setIterations] = useState<number | null>(null)
+  const [seed, setSeed] = useState<string | number | null>(null)
   const [optimizing, setOptimizing] = useState<boolean>(false)
   const [result, setResult] = useState<QueryResult | null>(null)
   const [error, setError] = useState<QueryError | null>(null)
@@ -45,6 +48,8 @@ const OptimizePage: NextPage = () => {
           seq: seq,
           org_list: species.map((s) => s.id),
           weights: weights,
+          seed: seed,
+          iterations: iterations,
         }
       )
       .then((res: AxiosResponse<QueryResult>) => {
@@ -128,7 +133,12 @@ const OptimizePage: NextPage = () => {
               Input sequence:
             </p>
             <SeqInput seq={seq} setSeq={setSeq} seqType={seqType} />
-
+            <HyperParameterSelection
+              iterations={iterations}
+              setIterations={setIterations}
+              seed={seed}
+              setSeed={setSeed}
+            />
             {optimizing ? (
               <button
                 className="px-4 py-2 my-2 text-lg font-bold text-white transition-all bg-blue-800 border-2 border-blue-800 rounded-lg cursor-wait"

web/yarn.lock

@@ -894,9 +894,9 @@ camelcase@^5.3.1:
   integrity sha512-L28STB170nwWS63UjtlEOE3dldQApaJXZkOI1uMFfzf3rRuPegHaHesyee+YxQ+W6SvRDQV6UrdOdRiR153wJg==
 
 caniuse-lite@^1.0.30001271, caniuse-lite@^1.0.30001272, caniuse-lite@^1.0.30001283:
-  version "1.0.30001312"
-  resolved "https://registry.yarnpkg.com/caniuse-lite/-/caniuse-lite-1.0.30001312.tgz#e11eba4b87e24d22697dae05455d5aea28550d5f"
-  integrity sha512-Wiz1Psk2MEK0pX3rUzWaunLTZzqS2JYZFzNKqAiJGiuxIjRPLgV6+VDPOg6lQOUxmDwhTlh198JsTTi8Hzw6aQ==
+  version "1.0.30001423"
+  resolved "https://registry.npmjs.org/caniuse-lite/-/caniuse-lite-1.0.30001423.tgz"
+  integrity sha512-09iwWGOlifvE1XuHokFMP7eR38a0JnajoyL3/i87c8ZjRWRrdKo1fqjNfugfBD0UDBIOz0U+jtNhJ0EPm1VleQ==
 
 chalk@^2.0.0, chalk@^2.4.2:
   version "2.4.2"